part 1] IGNEOUS AND ASSOCIATED ROCKS OF LLANW'KTID. 35 



The range in size of the spilite-blocks as seen in the hand- 

 ;specinien has already been noted. The larger blocks are identical 

 with the rock of the flows above. Under the microscope much 

 smaller fragments are to be seen, mixed with a few pieces of felsitic 

 xock and many broken quartz- and felspar- crystals. All these are 

 set in a murky brownish-black matrix, probably decomposed glass- 

 dust, but now quite unresolvable. 



While the spilite-fragments compare in general characters with 

 the lavas described below (p. 36), significant differences are not 

 wanting. It is important to note, for instance, that a good flow- 

 structure is common in the smaller fragments, but absent in the 

 spilite-flows. This points to a much lower viscosity, and suggests 

 that the fragments were derived from a rock which crystallized in 

 the deeper parts of some vent, where conditions of flow were 

 accompanied by that retention of fluxes which commonly charac- 

 terizes the hypabyssal and plutonic phases. As against this, how- 

 ever, is the fact that the spilite-fragments are decidedly of finer 

 grain than the overlying lavas, a fact not easily reconciled with 

 high flux-content and crystallization at relatively great depth. 

 Moreover, not a few of the fragments show a remarkable develop- 

 ment of small chlorite-filled vesicles. These vesicles make up at 

 least 50 per cent, by volume, of the rock, and impart to it a quite 

 remarkable appearance. This, again, points to high gaseous content 

 and uniform pressure ; but the fragments in question can hardly 

 have crystallized under the same conditions as those showing good 

 flow-structures, which are generally non-vesicular. 



The fragments are either ragged in outline (when non-vesicular), 

 or they exhibit beautiful bogen-struktur with the very charac- 

 teristic concave outlines of broken vesicles (PI. I, fig. 2). 



Secondary rearrangement of the matrix has produced peculiar 

 spherical bodies which present a pisolitic appearance. Certain 

 ■crystal-tuffs, described by Dr. Herbert H. Thomas from the Lower 

 Llanvirn Beds of the Carmarthen area, show this feature. A 

 possible phosphatic composition is suggested. 1 



Mr. C. I. Gardiner & Prof. S. H. Reynolds, in their paper 

 quoted above (p. 20) come to no decision as to whether the spilite- 

 breccia which they describe is a flow-breccia or an explosion- 

 product. There seems little doubt that the Llanwrtyd rock is of 

 the latter nature : that is, it is an agglomerate. There is, as we 

 have noted, considerable variation in the nature of the spilite- 

 fragments. Mr. J. F. N. Green 3 has remarked, with reference to 

 the autobrecciation of a submarine lava, that the 



' mixed mass . . . will form an intricate mixture of blocks, differing slightly 

 in vesicularity, crystallization, proportion of phenocrysts, etc., cemented by 

 similar material.' 



In this case the variation in the blocks is more than slight, and 



1 'The Geology of the South Wales Coalfield, pt. x : The Country around 

 Carmarthen ' Mem. Geol. Surv 1909, p. 35. 



2 Proc. Geol. Assoc, vol. xxx (1919) p. 161. 



D2 



