COABSELT SPHEKXJLTTIC ROCKS. 185 



than 1 centim., and in places a bare 3 millim. in thickness. The 

 hollow shell thus formed by decomposition is now entirely filled 

 with quartz ; and it is, indeed, questionable whether this relatively 

 large cavity ever existed as a whole, it being perhaps the result of 

 successive excavation and infiltration. 



Seen in section, the rock is often traversed by rudely parallel 

 cracks, some of which are posterior even to the infilling of the 

 hollows of the altered spherulites. These cracks, filled with chal- 

 cedony, are remarkably wide wherever they traverse spherulitic 

 matter, but become mere Hues, for the most part, while passing 

 through the matrix. Nor is this due to their being disguised by the 

 secondary devitrification of the glass *, since the perlitic curves 

 across which they occasionally run exhibit no such break as a wider 

 fissure would have caused (PI. IX. fig. 2). If, as is highly probable, 

 the variation in width of these cracks has been determined by decom- 

 position along their walls, their sudden expansion when they enter 

 even the smaller spherulites is somewhat striking evidence of the 

 greater stability of the matrix, and would lead us rightly to infer 

 that the agents which excavate cavities in such rocks would produce 

 a marked effect on the spherulites alone f. 



The rocks of Saxony and the Wrekin area prepare us, then, for 

 finding a mass composed of great nodules of quartz, each enclosed in 

 a felsitic envelope, the products of alteration being thus out of all 

 proportion to such primary stiuctures as remain. As we have on a 

 >imall scale spherulitic rocks in which the segregations outcrowd the 

 matrix, so we have in the Silurian rhyolite of Digoed J, near Pen- 

 machno, a mass some 15 feet thick, consisting almost entirely of white 

 close-set nodules. These reach 7 or 8 centim. (3 inches) in dia- 

 meter, and are almost all filled with quartz, the outermost layer of 

 the spherulite being often alone preserved. A chloritic mineral is 

 generally associated with the quartz (PI. IX. fig. 3). Here and there 

 the spherules suddenly diminish in size until their exceptional cha- 

 racter is lost, while the matrix for a short interval assumes its normal 

 predominance. Prof. Bonney §, to whose detailed investigation and 

 description of Xorth-Welsh nodular felsites students of this area must 

 always turn, has referred to this rock in connexion with similar 



* The product of this devitrification seems very largely to be lelspar, twinning 

 even being exliibited by many of the granules. Cf. Eonney, Proc. Geol. Soc. 

 1885, p. 94. 



t It may be worth while to state the exact position of the quarrj- where the 

 best specimens of the Lea Kock lavas seem always to have been obtained, since 

 there are few excations in the area. The road from Wellington joins the main 

 Shrewsbury road at a point called " Hay Gate" on the 1-inch Ordnance Map. 

 1^ mile west of this is a t()ll-house,»mai-ked T. P., on the left. A large new 

 red mansion lies a little further on upon the right, and beyond its lodge ie a 

 small cross road. The left liand or southern portion leads at once into the 

 quarry. The name Lea Rock does not now seem to be recognized in the district. 



I The farm of Digoed is f )und on the 1-incli Ordnance Map by following the 

 biffh road south-south-east from Ijettws-y-Coed till a branch goes off to Pen- 

 machno above the Conwy Falls. Digoed is set high on the cast side of this 

 branch, about 1 mile towards Penmachno, and the nodular rock rises in one or 

 two bosses near the house. 



§ Q. J. G. S. vol. xxxviii. p. 289. 



