150 Scientific Proceedings, Royal Dublin Society. 



granite may go on by ionic rather than by magmatic penetration. 20 Though 

 the transference is selective, it seems probable that continued interaction 

 wholly destroys the original nucleus, by the opposite diffusion of its materials 

 into the great mass of molten highly siliceous matter round about. Such 

 matter, as Shand points out, rises in temperature until solution of the more 

 basic substance is complete. While oligoclase and andesine are formed during 

 the earlier stages, which are recorded in the spherulites of orbicular rocks, the 

 presence of potassium felspar in an inclusion, where this mineral also exists 

 in the surrounding rock, marks an advanced stage of destruction. 



C. W. Drysdale 27 has recently recorded a remarkable instance of the 

 replacement of invaded material by an igneous rock. Further details would 

 seem desirable ; but it appears that " the more permeable pebbles and 

 matrix " of a conglomerate of Eocene or Oligocene age are now represented 

 by pseudomorphs of syenitic constitution. The syenite penetrated the con- 

 glomerate in Miocene times, before the stratum was consolidated, so that 

 the completeness of the interaction, rivalling that in composite gneisses, 

 cannot be ascribed to exceptional conditions of metamorphism in the lower 

 layers of the crust. 



These considerations render it less difficult to regard the granite nuclei in 

 the rock of Mullaghderg as of external and pseudomorphic origin, whether 

 they include a residue of schist or not. Similarly, the "granulite" nuclei in 

 the orbicular rock of Walkraal, described by W. J. Gau, 28 may represent the 

 almost complete assimilation of more basic inclusions. They are closely set, 

 and are surrounded by pegmatitic shells, and they differ from the medium- 

 grained hornblende-granite or granulite that surrounds them in containing 

 rather more biotite than hornblende. F. D. Adams and A. E. Barlow 19 have 

 critically considered the sheets of spherulites that occur in the granite of 

 Pine Lake, and have well compared them with axiolitic bands in lava- 

 flows. They are mainly composed of quartz, nmscovite, and sillimanite, 

 but contain a small amount of orthoclase. The wall-rock is " a basic gabbro- 

 like amphibolite," from which, as the authors point out, the spherulites are 

 not likely to have been derived. They regard the nodules as segregations of 

 more acid material from the magma, but suggest, however, that " schlieren " 

 may have been brought from another part of the igneous reservoir. May we 

 not see, then, in these spherulites the result of the enrichment in silica of 



26 Ibid., p. 131. 



27 "Geology o£ Franklin Mining Camp," Geol. Surv. Canada, Mem. 56 (1915), p. 82. 



28 " Notes on the Bush veld in the Neighbourhood of the Junction of the Elands and 

 Oliphants Rivers," Trans. Geol. Soc. S. Africa, vol. ix (1906), p. 70. 



20 "Geology of the Haliburton and Bancroft Areas, Ontario," Geol. Survey Canada, 

 Mem. 6 (1910), p. 127. 



