136 SANDSTONE BLOCKS IN GRAVEL OF SOMME. CHAP. vm. 



A stone hatchet of an oval form, like that represented at 

 fig. 9, p. 115, was discovered at the same time, about one foot 

 lower down, at c, in densely compressed gravel. The surface 

 of the fundamental chalk is uneven in this pit, and slopes 

 towards the .valley-plain of the Somme. In a horizontal 

 distance of twenty feet, I found a difference in vertical height 

 of seven feet. In the chalky sand, sometimes occurring in 

 interstices between the separate fragments of flint, constituting 

 the coarse gravel No. 4, entire as well as broken fresh-water 

 shells are often met with. To some it may appear enigmatical 

 how such fragile objects could have escaped annihilation in a 

 river-bed, when flint tools and much gravel were shoved 

 along the bottom ; but I have seen the dredging instrument 

 employed in the Thames, above and below London Bridge, 

 to deepen the river, and worked by steam power, scoop up 

 gravel and sand from the bottom, and then pour the contents 

 pell-mell into the boat, and still many specimens of Limnea, 

 Planorbis, Paludina, Cyclas, and other shells might be taken 

 out uninjured from the gravel. 



It will be observed that the gravel No. 4 is obliquely stra 

 tified, and that its surface had undergone denudation before the 

 white sandy loam, No. 3, was superimposed. The materials 

 of the gravel at d must have been cemented or frozen together 

 into a somewhat coherent mass to allow the projecting ridge, 

 dy to stand up five feet above the general surface, the 

 sides being in some places perpendicular. In No. 3 we 

 probably behold an example of a passage from river-silt to 

 inundation mud, or loess. In some parts of it, land shells 

 occur. 



It has been ascertained by MM. Buteux, Ravin, and other 

 observers conversant with the geology of this part of France, 

 that in none of the alluvial deposits, ancient or modern, are 

 there any fragments of rocks foreign to the basin of the 

 Somme no erratics which could only be explained by sup- 



