318 W. F. SMEETH. 



it seems to me that they differ also in two important particulars: 

 First, they do not lie in the meshes of a set of polygonal cracks, 

 but on the contrary are often seen to be continuous with some of 

 the polygonal cracks of the matrix. Second, instead of being 

 segmental or touching each other tangentially they are almost 

 invariably closed by abutting abruptly against each other. 



These I consider to be important distinctions, and I cannot 

 imagine anyone, who can view the two sets of cracks side by side, 

 calling them by the same name, save on an extremely general 

 basis of classification. 



As I have not been able to obtain a photograph exhibiting the 

 perlitic cracks and those in the quartz side by side, I have 

 endeavoured to reproduce in Canada balsam a structure similar in 

 essential points to that of the rock. The result is shown in 

 [Plate 45, fig. 4) and was obtained as follows : — 



Small round pieces of polished cover glass were cemented on to 

 a ground glass, a layer of balsam was spread over this and heated 

 to a certain point and suddenly cooled. The result, as shown, 

 was polygonal and perlitic cracks in the balsam overlying the 

 ground glass and curved and sometimes concentric cracks in that 

 overlying the polished pieces. It will I think be admitted that 

 these latter cracks resemble in character those in the quartz grains 

 and that they are quite distinct from the undoubted perlites. 



There are a few other points in connexion with this matter to 

 which I must allude. If a curved crack which lies mainly in the 

 quartz encloses also a portion of the glass, one or more concentric 

 cracks, representing an imperfect perlite cut off by the quartz, 

 may be present in the enclosed glass. Again, if a perlite is 

 adjacent to a crystal one or more of the cracks may traverse the 

 crystal (Plate 45, fig. 1). Or, if a small crystal occupies the centre 

 of a perlite, one or more of the cracks may traverse the crystal. 

 These cases however do not prove that perlitic cracking will 

 occur in the crystal per se. 



There is another case which is probably of frequent occurrence 

 and of which I have an example beside me at the present moment 



