OF THE ROSS ARCHIPELAGO 139 



and in each case run parallel to the dominant faces, i.e., the cleavage net is respectively 

 rectangular or triangular, and therefore probably dodecahedral. The refractive index, 

 superior to that of xylol (/* = 1 '494), excludes analcite and sodalite, so that the mineral 

 must be leucite. This too is what would be expected in the melt of a rock consisting 

 largely of sanidine. 



There is no clue as to the rock in which this inclusion was remelted, but the above 

 description renders it probable that the sanidinite originated in a trachyte. 



Origin of the Sanidinites and Microtinites 



The origin of sanidinites in trachytes appears at first sight very simple. They consist 

 in general of the same minerals as the phenocrysts of the rock, and appear to have 

 arisen by a simple aggregation of these minerals; another plausible view is that 

 they represent a complete crystallisation in depth of a portion of the trachytic magma, 

 perhaps as a wall to the magma basin, and have been broken off and included in the 

 still unconsolidated magma on eruption. Lacroix, in his earlier work,* considered the 

 latter the more probable mode of origin ; but subsequently he has shown that sani- 

 dinites may arise in many different ways. For instance, trachytes appear to have the 

 power of converting inclusions of such rocks as older trachytes, gneisses and granulites 

 into sanidinites | and this may have happened in many cases where the intermediate 

 steps can be no longer traced. A still more curious origin is put forward for some 

 sanidinites in the basic leucitic tuffs of Somma.J Besides the sanidinites the tuffs 

 contain blocks of metamorphosed limestones which have locally the constitution of 

 sanidinites, and the isolated sanidinites are assumed to be fragments broken off from 

 the limestone blocks. The origin of the sanidine in the cavities of the limestone is 

 ascribed to contact alteration accompanied by an exudation of soda into the limestone 

 under pneumatolytic conditions. It is obvious, therefore, that the explanation of the 

 origin of the sanidine in any rock must be carefully scrutinised. 



The peculiar feature of the Mount Cis sanidinites is the predominance of olivine 

 over »girine-augite, in opposition to its rarity in the trachyte. A further mineralogical 

 difference between the two rocks is the absence of the alkaline amphiboles in the former 

 and their presence in the latter. Chemically expressed, the sanidinites are richer in iron 

 and magnesia than the trachytes and poorer in lime. They cannot therefore be regarded 

 as simply a deep-seated crystallisation of a part of the trachyte magma, unless some 

 differentiation is admitted. One specimen might suggest that they are simply aggrega- 

 tions of phenocrysts (Fig. 4, PI. II), but phenocrysts of sanidine are not at all common 

 in the trachytes, and the specimen might be interpreted equally well as a sanidinite 

 that had been partly broken up by movement in the trachyte. § 



The microsanidinites present the same mineralogical differences from the trachytes 

 as the sanidinites. From their orthophyric structure they point more clearly to deep- 

 seated crystallisation, with a slight amount of differentiation. The development of 

 a thin zone of coarse crystals around one of them (1753) perhaps gives the clue to the 

 origin of the sanidinites. The latter may have been originally inclusions of orthophyric 

 trachyte which the trachyte magma has by the power of its mineralising vapours trans- 



* Les Enclaves, etc., pp. 353 et seq. 



•f- Lacroix, A., " Sur deux nouveaux groupes d'enclaves des roches eruptives," Bull. Soc. fr. 

 Min., xxiv, 1901, pp. 488-501. 



J hoc. cit. 



§ This specimen was sent away to the International Geological Congress at Stockholm, so that 

 the writer has not been able to re-examine it. 



11 u 2 



