Vol. 59.] IN GLASSY IGNEOUS ROCKS. 439 



conditions of solidification must have given one of the constituent 

 minerals, commonly the felspar, a very slight advantage over the 

 other, and this structure generally belongs to both microgranites 

 and ordinary granites. The curvilinear type, however, is seen in 

 many compact felstones and in a large number of granitoid rocks, 

 especially in that group (generally of pre-Cambrian age *) which 

 were formerly supposed to have undergone metamorphism. I 

 regard these curvilinear boundaries as signifying that the tempera- 

 ture remained steady at a height which, in a normal granite of 

 corresponding coarseness, would have allowed felspar to separate 

 from the residual quartz in the usual way, but that owing to some 

 disturbing factor (probably an increased viscosity) the latter mineral 

 had a greater power of resistance ; so that a rectilinear boundary 

 was impossible, because neither could definitely overcome the 

 other. Thus, as we shall presently see, the significance of curvi- 

 linear boundaries to grains in a granular rock is very similar to 

 that of root-like or wavy structure as opposed to rectilinear in a 

 ' graphic ' rock. 



I do not forget that in noncrystalline igneous rocks a granular 

 structure has been attributed to movement in the act of cooling, 

 which has rubbed off the angles of crystals already formed by the 

 resistance of the viscous residue. This is obviously a possibility, 

 and I believe that conspicuous crystals are sometimes thus treated, 2 

 but we must remember that the residual magma which this expla- 

 nation requires to be present would either crystallize independently 

 as a groundmass (as in the cases just mentioned), or would be used 

 up in augmenting, and thus repairing, the damaged crystals. 



In the more compact felsitic rocks we not unfrequently find a 

 slightly- variable structure, some patches being a little coarser than 

 others — spherulites being restricted to particular bands, etc. : see 

 PI. XXVI, fig. 3. These may be attributed to slight local variations 

 in chemical composition, with which we are familiar in ordinary 

 fluxion-structure. 3 



Some compact felstones, when examined under the microscope, 

 exhibit a close association of spherulitic, micrographic, and granular 

 structures (the second being often root-like, and the third curvilinear 

 in outline), the one apparently passing almost insensibly into the 

 other : see PI. XXVI, figs. 4, 5, & 6. The relation of the first to the 

 second has already been noticed, and that which it bears to the third is 

 shown by studying sections. As these become more nearly tangential, 

 the radial is replaced by the granular structure, and the more crowded 



1 In these, as I have more than once pointed out, the structure often closely 

 resembles that of a quartzite. 



2 This, as I have explained in Quart. Journ. Geol. Soc. vol. xlvii (1891) 

 pp. 483-90, accounts for an occasional ' augen-structure ' in rocks wherein 

 there is evidence of tension but none of pressure. 



3 Something of the kind also occurs, as has been frequently noticed, in holo- 

 crystalline rocks. Often it is indicative of a partial melting-down of one rock 

 by another, but sometimes is more suggestive of differentiation (prior to the 

 movement) in a magma. 



