MELAPIIYRES AND FELSITES OF CARADOC. 537 



their bouudaries, such as arc invariably seen on the surfaces of 

 the fragments of tuffs formed from vesicular or scoriaceous lavas. 

 (Pl.XIX. lig. 4.) 



The general aspect of the hand-specimen is that of a dark iron- 

 grey or bluish-grey rock, compact iu texture, but ap])earing slightly 

 vesicular when examined under a pocket-lens. It also shows some 

 irregular spots of yellowish-white to })inkish-white crystalline matter, 

 which effervesces briskly when touched with a drop of acid. 



iu thin section, under the microsco])e, the fragments composing 

 the rock are seen to vary considerably in translucency, the ground- 

 mass in some being a more or less completely devitrified pale-brown 

 glass, containing opaque matter which appears merely as fine dust 

 under a magnifying power of 250 linear. In other fragments the 

 opacjue particles in the groundmass are larger, and the rock of which 

 these fragments consist seems to differ in no appreciable respect 

 from the mclaphyre previously described which occurs in the imme- 

 diate vicinity of this tuff. 



In many of these fragments the magnetite grains are seen to b(^ 

 more closely massed along the margins of the fragments, giving rise 

 to a narrow and perfectly-opaque black border, while in some 

 instances the entire groundmass of the fragment has been rendered 

 absolutely opaque through development of magnetite. 



It has already been shown that a basic, vitreous lava, such as the 

 basalt-glass of Kilauea, when heated for 260 hours at a temperature 

 ranging from 700*^ to 1200° C, becomes strongly magnetic and 

 absolutely opaque through separation of magnetite *. Bearing this 

 fact in mind it may, I think, be inferred that the fragments con- 

 stituting the mclaphyre tuff of Caer Caradoc have not resulted from 

 the mere crushing of a lava, but that they were ejected from a crater 

 as lapilli and volcanic sand ; that their surfaces, in many instances, 

 were sufficiently heated to give rise to the formation of an opaque 

 superficial crust of magnetite, while, in other cases, a more protracted 

 roasting carried this process to its extreme limit, so that all the 

 iron present in the lava in the protoxide state became converted 

 into the magnetic oxide. 



The action exerted upon a magnetic needle, when a specimen 

 of the melaphyre lava (No. XII.) is brought near it, is exceedingly 

 slight compared with that produced by a considerably smaller 

 specimen of the mclaphyre tutt' (Xo. XIV.). Pig. 4 in PI. XIX, 

 represents part of a section of this tuff, magnified 18 diameters. 

 The fragments and portions of fragments here shown exhibit a 

 perfectly opaque groundmass, while the small felspar- crystals and 

 skeletons lying in this groundmass appear unaltered and remain 

 perfectly translucent. The broad light band passing diagonally 

 across fig. 4 represents cementing material which, in this rock, 

 consists of chalcedony. Irregularly shaped cavities occur in this 

 cement, and these have been filled partly with pale green serpen- 



* * Notes on Alteration induced by Heat in certain Vitreous Rocks,' Proc. 

 Boyal Soc. vol. xl. (ISSfi) p. 437. 



