OCCURRENCE IN ANCIENT BRITISH LAYAS. 167 



lens ; and there can be little doubt that the original " Lithophysen " 

 are records of exactly the same action. It is interesting to find 

 very similar appearances in our ancient rhyolites of the Wrekin, 

 the roughly concentric cavities formed by the weathering of large 

 spherules becoming lined with crystallized quartz (PI. IV. fig. 4). 

 The red earthy fibrous centres of the ordinary Wrekin spherulites *, 

 contrasting with their clearer borders, are probably similar evidence 

 of the more ready decomposition of the radial and perhaps toore 

 basic portions. 



The one difficulty of this view of the formation of hollow 

 spherulites is the getting rid of the kaolin that will, in all proba- 

 bility, be formed. But. even if the alumina is not removed in soluble 

 combinations, the extreme minuteness of kaolin-particles will 

 enable them to enter and pass along the cracks of a vitreous lava 

 with comparative ease. Perlitic fissures, to which the more 

 obvious joints of a hand-specimen are veritable gullies, may at last 

 become choked, but are none the less active as transporting channels. 

 Some of them, in an interesting specimen brought by Mr. Car- 

 ruthers from the Yellowstone area, measure from -02 to -06 millim. 

 across ; and the sharpness of their edges removes the suspicion that 

 they may have been much widened by decomposition. Now large 

 particles of kaolin, such as one may rub off altered felspar with the 

 finger, may measure -01 millim. and more ; but the material 

 washed down naturally from a granite area is seen to consist of 

 specks far smaller, -004, -002, -0015 millim. or less in diameter. The 

 way in which kaolin-dust penetrates aU the cracks of a decomposing 

 granite, and even the planes of separation between different minerals, 

 is fair evidence of its capacity for wandering from its parent source ; 

 and the frequent occurrence of "Lithophysen" in rocks showing 

 perlitic structure may be due to the greater number of channels 

 there provided through which the alteration-products may run 

 their course f. 



To return at last to the Silurian felsites of the Llanberis Pass, we 

 find in microscopic sections ample proof of their spherulitic character 

 and of the community of origin of both solid and hollow nodules 

 (PI. lY. fig. 5). The relics of radial structure are faint but indubi- 

 table, and the junction of the spherules and the glass is usually well 

 defined. Quartz may be seen developing among the crypto- crystalline 

 hbres, and brilliantly surrounding the central hollow when it occurs. 

 But in place of this cavity in several cases are the remains of original 

 smaller spherulites, apparently of more complete development than 

 the surrounding felsitic nodule, and therefore more liable to decay. 



* S. Allport, " On certain ancient devitrified Pitchstones and Perlites from 

 Shropshire/' Quart. Journ Geol. Soc. vol. xxxiii. p. 454. 



t I cannot refrain from here referring to a case bearing, pei'haps distantly, 

 on hollow spherulites. In much-altered granite (luxullianite) boulders on the 

 slopes above the inlet of Nanjisal, near the Land's End, many of the felspar- 

 crystals are found to be nothing but hoUovp shells lined with well-developed 

 quartz. In places this infiltered silica has entirely replaced the orthoclase 

 and formed solid pseudomorphs that are completely deceptive at a distance. 



