236 ME. H. H. THOMAS ON THE [Aug. 1909, 



extremity of Torbay, and associated with the breccias of Dawlish 

 and Heavitree (near Exeter). 



The cementing-material in the Lower Breccias is almost always 

 of a ferruginous nature, only becoming calcareous in the neighbour- 

 hood of limestone-masses of Devonian or Carboniferous age. A 

 complete silica-cement never occurs ; but secondar}- growths of 

 silica have been noted in a few cases (as, for instance, Berry Head). 

 Lithologically, the Lower Sandstones are similar to the beds of sand 

 and sandstone associated with the Lower Breccias, and the same 

 remark applies to the cementing materials. 



The particles are all well rounded ; but the rounding is most 

 complete in the centre and south of Devon, especially in the Broad- 

 Clyst district. Around the Somerset inliers of older rocks the 

 rounding is not so well marked. 



The grains forming the marls below the Pebble-Bed, owing to 

 their extremely small dimensions, cannot readily be compared with 

 those of other New Bed divisions. The material, as might be 

 expected from the size of the particles, is all angular ; but, when it 

 becomes sufficiently coarse to form beds of sand or sandstone, the 

 grains are rounded. Well-rounded sands occurring in the Lower 

 Marl subdivision have been noticed at St. Mary's Clyst, Hulham 

 (near Exmouth), and several other localities ; but the rounding is 

 never so complete as in the sands of the lower divisions, and 

 approximates more to that observable in the grains forming the 

 matrix of the Pebble-Bed. 



The finer grains of the Pebble-Bed are remarkably constant in 

 character: they are subrounded, after the manner of fluviatile or 

 marine sands. The constituents of the matrix of the Upper Marls 

 and Sandstones, and of the Upper Keuper Sandstones of the 

 Bridgwater area, are all angular to subrounded, according to size. 

 Generally speaking, the rounding of the grains is less marked in 

 this division of the New Red than in the sands of any other 

 division. 



With regard to the cementing-materials in the Upper Marls and 

 Sandstones, in addition to the usual ferruginous material we find 

 that calcium-carbonate is almost universal, and that substances 

 such as barytes and gypsum play an important part locally. 



In the marls, often far removed from airy limestone-outcrop, 

 the percentage of carbonate is extremely high : for instance, four 

 samples taken from the deep boring at Lyme Begis 1 yielded the 

 following percentages of carbonate, estimated as CaCO^ : — 



Depth of 437 feet 11-51 



850 „ Kv49 



„ 1200 „ 21-10 



„ 1300 „ 12-26 



Barytes was a prevalent constituent of a sample taken from a 

 depth of 1000 feet, and undoubtedly existed as a cement. 



1 A.J. Jukes-Browne, Quart. Journ. Geol. Soc. vol. hiii (1902) p. 284. 



