One of the best sections in the district for the exemplification of the internal structures of these lavas is that in the large quarry at the top of Posbury Hill. On the west side of this quarry the rock is tolerably compact, but contains vesicles and irregular steam-holes. On the east side it passes upward and laterally into a coarse agglomerate of its own fragments, and in its mass it encloses similar agglomerate. No sharp passage can be traced between the two rocks. So far as I could judge, it seemed to me that the lava had broken up as it moved along, possibly shattered by coming in contact with water. The agglomerate is overlain by some reddish ashy sandstone, which fills up the interstices between the slags, and is immediately covered by a bed of lilac andesite, marking another distinct outflow. [Illustration: Fig. 231.--Section of agglomerate overlain with sandstone and andesite, Posbury, Crediton. ] As in Ayrshire, the lavas of Devonshire are not accompanied by any thick accumulation of tuff. The fragmentary discharges consisted in both areas of fine dust and gravelly detritus of small lapilli, which were not ejected in such quantities as entirely to conceal the ordinary non-volcanic sediment of the water-basin. The dust and cinders mingled with the red sand and angular scree-material, so that we now see a group of red, somewhat ashy sandstones and breccias. Among the component fragments of the breccias, a considerable variety of igneous material may be observed. While the most of the non-volcanic stones may have been derived by ordinary processes of weathering from rocks exposed at the surface, it is by no means improbable that some of them, including even pieces of Culm grit, killas and baked slate, may have been ejected from volcanic vents.[111] [Footnote 111: On the composition of the Devonshire breccias see Mr. R. N. Worth, _Quart. Journ. Geol. Soc._ vol. xlvi. (1890), p. 69. This author has adopted the view that the granite of Dartmoor represents the neck of a great volcano from which these later volcanic materials were ejected. But all the evidence seems to me in favour of numerous small vents situated not far from the outcrops of the lavas, as stated in the text. See Mr. B. Hobson, _Quart. Journ. Geol. Soc._ vol. xlviii. (1892), p. 498. The Dartmoor granite is later than the surrounding Carboniferous rocks, but no good evidence has been obtained to connect it with the Permian volcanic phenomena of Devonshire.]