Vol. 56.] CEYLON KOCKS AND GRAPHITE. 607 



and pyroxene. The constitution of the viscous magma would be 

 altered by the absorption of garnet-material, which would also tend 

 to increase the viscosity in the neighbourhood of the garnets. So 

 that in the area between the now corroded garnets there was present 

 a pasty felspathic magma containing in solution the remnants of the 

 pyroxene-elements, with some derived from the corroded garnets. 

 As cooling proceeded some pyroxene was added to the original grains 

 (the appearances tend to suggest this), and the remaining narrow 

 strips of magma consolidated as an intergrowth of pyroxene and 

 felspar, the felspar forming ophitic plates and the garnets serving 

 as a point of attachment for the root-like fingers of pyroxene. 

 Thus, in the end, we have slightly more pyroxene present than if 

 the garnets had not been corroded, for portions of the magma 

 which would have consolidated as pnre felspar may, with the addi- 

 tion of magnesia, iron, and calcium which a pyrope would supply, 

 have actually formed a pyroxene-felspar intergrowth. Inter- 

 growths of plagioclase, pyroxene, biotite, and magnetite in connexion 

 with garnets occurring in a llagedara rock are somewhat similar ; 

 and indications of the same structure are found in a granulite from 

 Talatuoya. These radiating coronas differ from those described by 

 Prof. Lacroix in the root-like form of the pyroxene. 



In the pyroxene-granulite of Ragedara, associated with the 

 graphite (see p. 594, text-fig. 1, & PI. XXXIII, fig. 2), somewhat 

 similar phenomena seem to have occurred; but here the prevailing 

 felspar is orthoclase, and accordingly the result of the corrosion of 

 the garnets has been the formation of brown mica. Probably the 

 garnets appeared earl}' in the history of the rock, and served as 

 nuclei round which the felspathic material tended to collect : sub- 

 sequent change in the physical conditions caused the corrosion of 

 the garnets, without sufficient movement taking place in the viscous 

 magma to separate the products of corrosion from the garnets. 



The occurrence of centric and granulitic structures in pyroxene- 

 granulites was noted by Dr. Becke ; and it probably has some special 

 significance in connexion with their origin which is not yet fully 

 understood. 



The contemporaneous formation, by the interaction of two 

 minerals, of a third mineral or mineral variety is comparable with 

 the fact that in any ordinary rock in which, for example, hornblende, 

 biotite, and felspar are found, the biotite is often intimately asso- 

 ciated with the hornblende. This probably results from a slight 

 mixture of materials at the edge of the newly-formed hornblende- 

 crystals, taking place as the magma consolidates. 1 



(16) Microperthite. 



The oi\-io_Iase of Ceylon rocks is scarcely ever without the 

 fibrous microperthitic structure. In many cases it can be seen, even 



1 [See Parsons, ' The Development of Brown Mica from Augite by Eeaction 

 with Felspathic Material' Geol. Mag. 1900, p. 316.— July 15th, 1900.] 



2s2 



