Vol. 51.] THE COUNTRY AROUND FISHGUARD. 173 



larly marked where the centres of the meshes of the network have 

 no central dark dot. In cases of extreme silicification and recrystal- 

 lization these minute spheres have fused together and produced 

 crystalline patches of quartz extinguishing simultaneously, but the 

 centres and occasionally the outlines or the positions of the separate 

 spheres may in a few cases be detected by means of the persistent 

 radial inclusions. (PL VI. fig. 1.) 



The original groundmass is of a pale greenish or pale brownish 

 colour, and is isotropic or felsitic. The secondary quartz-mosaic 

 (not micropoikilitic) varies in irregularity of grain with the degree 

 of alteration. The perlitic and other structures, plainly shown in 

 some sections [316], [319], are scarcely discernible where the altera- 

 tion into the mosaic is extreme [320]. 



With regard to the origin of the ' minute spheres,' I think we 

 must regard them as simply very minute spheroidal structures 

 produced at the same time as the larger spheroido-perlitic structure, 

 for they graduate insensibly, when clustered together, into the 

 ' crackled china ' structure. Moreover, they tend to be obliterated by 

 the replacement of the rock by the quartz-mosaic aggregate. It 

 does not seem possible to regard them as spherulites of any kind, 

 though it must be granted that their origin is at present obscure. 



Another rock from the same locality — Bwlch Mawr — shows an 

 entirely different method of replacement, and, judging from its 

 peculiar features, was most probably a perlite with lithophysse. In 

 a section of it [83] we see irregular curving shreds and bands of 

 clear quartz ; sometimes a long, curved, jagged-edged band behaves 

 as a single crystal with crossed nicols. But in ordinary light perlitic 

 structure can be seen, being indicated by the more transparent 

 curving bands and lines of quartz. The abrupt way in which these 

 bands terminate, and their more or less sporadic and disconnected 

 manner of occurrence, point to a shattering of the rock after the 

 formation of the lithophysse and perlitic curves, and an irregular 

 opening of these cracks at the same time into which silica infil- 

 trated, subsequently cementing the rock together again. In relation 

 to those quartz-bands or infilled cracks, or quite independent of 

 them, are circular or oval patches of cryptocrystalline silica which I 

 hold occupy the original lithophysal cavities. In places these 

 patches seem to consist of a minute micropegmatite or of a micro- 

 crystalline aggregate. They are frequently encircled by the curving 

 quartz-bands. The patches may show a faint radial structure, 

 but are not definitely recognizable pseudo-spherulites or incipient 

 true spherulites. Patches and angular grains or groups of grains 

 of clear quartz abound ; and lines of minute greenish granules some- 

 times help to define the perlitic cracks. 



All the peculiar features of this rock seem explicable on the 

 assumption of a primary shattering, followed by complete replace- 

 ment by cryptocrystalline and crystalline silica. 



In many of the felsites of this area vesicles are prominent 

 features, and are of interest as showing both the primary and 

 Q. J. G. S. No. 202. o 



