EXCAVATIONS AT DEWLISH. 215 



occupies the more superficial parts of the trench connecting 

 these pot-holes ? 



The extensive opening up of the deposits lately under- 

 taken indicates that there is little difference between the 

 mammaliferous part and the sand below, and that they 

 probably all belong to a single period. This, as far as the 

 meagre evidence goes, is somewhere about the date of the 

 Cromer Forest-bed, or latest Pliocene. 



The probable story of the infilling of the trench and pot- 

 holes is as follows. The original joint crossing the stream- 

 bed was gradually widened and its solution-pipes were en- 

 larged by the swirling water descending into the large pot- 

 holes we now see. Gradually erosion and solution opened 

 up wider underground channels in the chalk ; the water 

 got away more freely, and the stream became a winter- 

 bourne, sinking into pot-holes at various points along its 

 course, sometimes disappearing entirely and leaving the 

 deserted trench and pot-holes open, vertical-sided, and 

 free from water. Then the open pot-holes would act as 

 natural pitfalls into which the animals which passed up 

 and down the valley might fall. This we think was the 

 mode by which the large unworn elephant-bones got into 

 the fissure. The bones of these animals were quietly let 

 down and never afterwards disturbed, though perhaps the 

 small bones were removed by carnivora. The large bones 

 show no sign of having been brought to the spot by water, 

 and there is no evidence that the Devil's Brook was any- 

 thing but a small stream, such as it is now. 



The swirling action of the water laden with chalky sand, 

 as it sank in whirlpools into the pot-holes, would polish the 

 flints in the way we see. The abundant sponge spicules 

 would help to smooth the surfaces, and fine chalk-mud (i.e., 

 whiting) is just the substance used by a lapidary for the 

 final polish. 



This we think is the story of the Dewlish trench ; but 

 though we have cleared up some points, we feel that, as is 

 usual with any scientific enquiry, we have raised more 



