E. C. Martin — Neiv Red Gravels of the Tiverton District. 157 



Gasteropod Euomplialus vermis j Whidb., is also common in the gravels 

 and Pilton Beds. Bellerophon suhglolatus, M'Coy, is a characteristic 

 fossil in the Marwood Beds of the Barnstaple district ; it is also 

 common in the grits that occur further east near Wiveliscombe, and 

 in the grit fragments found in the gravels. 



At Chimney Down, in the north-east of the district, fossiliferous 

 fragments of Upper Devonian sandstone and grit are extremely 

 common, whilst trap fragments appear to be absent. This fact, as 

 well as the wide distribution of the fossiliferous fragments, completely 

 disposes of the Bev. W. Downes' theorj^ of volcanic ejection. The 

 facts tend, rather, to prove a drift from the north-east during the 

 period of dejposition of the gravels. The grit and sandstone matrices 

 suggest a derivation from the Wiveliscombe end of the Pilton Beds 

 rather than from the more slaty beds found further west. The 

 Devonian fragments in the gravels may even have been derived from 

 a still more easterly extension of the Pilton Beds, now buried under 

 Triassic deposits. 



The angular and subangular nature of the fragments composing the 

 gravels, as well as the non-separation of coarse and fine material, 

 gives the impression that the component fragments have not travelled 

 far, and yet at White Down Copse they are found at least 15 miles 

 from the outcrop of the Pilton Beds. The fossiliferous pebbles found 

 at White Down Copse were all subangular. These JS^ew Bed deposits 

 certainly bear a considerable resemblance to the Boulder-clays, but 

 in the absence of striated pebbles (and none have yet been found) the 

 theory that they are glacial deposits is hardly admissible. 



Mr. Ussher regards the gravels as the result of torrential action 

 periodically operating on the margin of an area of depression attended 

 by periodic dessication and wind-drift. This supposition is in accordance 

 with our present views of the climate of England in Permian and 

 Triassic times, and would explain the coarse unstratified natui'e of 

 the deposits. The derivation from the north-east is natural on this 

 supposition, but difficult to explain on any other theory such as that 

 of marine agency. Eor the New Bed rocks are overlapped on the 

 margin of the Middle Devonian rocks near Williton, and thence to 

 Porlock Triassic rocks alone represent the series. Hence, on the 

 marine derivation theory a drift from the north is highly improbable, 

 and we would have to suppose drift from the east where the extension 

 of Devonian rocks may be expected to occur beneath the New Bed 

 rocks. The theory of torrential derivation from the north-east is 

 therefore most probable and is consistent with the appearance of the 

 deposits. These torrents carried the fragments as far as the entrance 

 of the Crediton valley, where they were met by other torrents 

 bringing down fragments of trap from the upper parts of this valley. 

 The fact that the Devonian fragments are apparently absent to the 

 west of Thorverton and that the gravels are replaced by a breccia 

 composed largely of trap is thus explained. 



The author's best thanks are due to Mr. J. T. Underhill for 

 proposing the investigation and to Mr. W. A. E. Ussher for some of the 



