214 EXCAVATIONS AT DEWLISH. 



pipes following the line of an east and west joint in the Upper 

 Chalk, But the pipes were subsequently greatly enlarged 

 by mechanical means, evidently by the swirling action of 

 the water charged with sandy gravel. This gravel was partly 

 derived from the flints removed from the surrounding chalk 

 during the scouring ; but it appears in the main to be com- 

 posed of broken and much weathered or waterworn flints, 

 such as will be found in the channel of any small chalk- 

 stream. The polishing, however, was done on the spot, 

 and is quite unlike anything found in the stones of the gravel 

 of an ordinary stream. 



The story told by the Dewlish deposits seems to he some- 

 what as follows. The Devil's Brook, when it flowed about 

 90ft. above its present level, perhaps only as a winterbourne, 

 met, at right angles to its course, an open joint, along which 

 pipes had already been formed by the percolating rain-water. 

 At this point the Brook sank into the chalk, the swirling 

 water transforming the pipes into pot-holes, which tended 

 to be continued downward till they reached the 

 saturation-level. 



It must not be thought that the present surface-level 

 around these pot-holes is any indication of the level at which 

 the river flowed. A glance at the bluffs which face the 

 Devil's Brook shows that here, as usual in the Chalk, there 

 were exceptional rushes of water during the Glacial Period, 

 when the soil was frozen. The valley has been deepened 

 and the bluffs rendered steep and precipitous since the 

 Dewlish pot-holes were cut, and it no longer shows the regular 

 curves, such as gentle action of the Brook would form. 

 Exactly how much the valley has been deepened since Newer 

 Pliocene times is not quite clear ; but the old river must 

 have flowed at the highest level at which the peculiar deposits 

 are found, and that is about 90ft. above the present level 

 of running water. 



. Two points remain : What was the age of the chalky 

 sandy gravel filling the pot-holes ? and what was their 

 relation to the overlying mammaliferous deposit, which 



