354 Becker— Determination of Plagioclase Feldspars. 



These facts clearly show that even the residual mother liquor 

 of a consolidating lava is far from homogeneous, even over the 

 area of a square centimeter, and therefore, also, that diffusion 

 in such a liquid must be extremely slow, since diffusion would 

 bring about homogeneity. The bearing of this observation on 

 the theory of " differentiation" of rock magmas is manifest. 



The existence of rod-shaped microlites, in connection with 

 the fact that some of them show Carlsbad twinning, throws 

 some light on the origin of twinning. If Carlsbad lamellae of 

 bacillar shape developed in contiguity but independently, they 

 would form X-shaped crystals. These are said to occur, but 

 are not common in my experience. * In the bacillar microlites 

 under discussion one lamella has controlled the development of 

 the rod, and the other lamella has accommodated itself. It is 

 difficult to see how this could happen in mere crystallization, 

 for why should one part resign its tendency to elongation in 

 the direction of the cleavage edge ? On the other hand, it is 

 well known, from the investigations of Messrs. Max Bauer, (X 

 Miigge, L. van Werweke, J. W. Judd, and others, that there is 

 much reason to ascribe polysynthetic structure largely to stresses 

 called-in play by the cooling process. This theory would fully 

 explain the bacillar microlites as well as some beautiful cases of 

 bent phenocrysts where twinning by various laws stands in 

 manifest relation to flexure. It would also afford an explana- 

 tion of the fact, of which comparison has convinced me, that 

 twinning is less frequent in the microlites than in the pheno- 

 crysts, for the smaller the crystal the less is the chance that 

 external stresses upon it would reach the intensity needful to 

 produce mechanical twinning. 



In the same connection it may be noted that stresses such, as 

 would lead to the mechanical twinning of plagioclases would 

 also set up a mechanical seolotropy in glass. As a matter of 

 experience, I have found it extremely difficult to detect any 

 absolutely isotropic base in some lavas which. I had reason to 

 believe contained a small amount of glass. The faint variation 

 of tints observed with the gypsum plate between crossed nicols 

 in such cases may, perhaps, have been due to strain and not to 

 crystalline symmetry of structure. 



* That phenocrysts often take the form of fully developed Carlsbad twins is 

 well known. 



