TYLOR AMIENS GRAVEL. 11 •") 



the chalk is convex on the average at L M, and concave on the ave- 

 rage at N Q. This, however, requires more explanation. 



The surface of the chalk between the line of 200 feet and of 120 

 feet above the sea is convex on the line N P, at the maximum 

 to the extent of 14 feet out of 80 feet perpendicular height ; and 

 we see very little gravel reposing on the convex surface. On 

 the contrary, in L M, between the 200-feet and 135-feet levels, the 

 maximum concavity is 15 feet ; and the great mass of the St.-Acheul 

 gravel is deposited in this hollow. 



But when we examine the surface of the chalk between and P, 

 between the 120-feet and 60 -feet levels, we find the chalk surface is 

 concave to the extent of a maximum of 29 feet, out of a total of 

 60 feet ; and, singularly enough, this 29 feet is almost exactly the 

 maximum thickness of gravel and loess in the great pit at Montiers, 

 where a section several hundred yards long is exposed. 



On the contrary, between the 130-feet and 76-feet levels on the 

 lines L M and I K, where the surface of the chalk is convex, there 

 is no gravel of any importance. 



In the section (Plate IV. fig. 1) between LM and IK the 

 chalk is nearly a straight line, falling 2|° between the 130-feet 

 and 90-feet levels ; we have 9 feet of gravel and loess exposed in 

 this favourable position for its accumulation. 



When we see the gradual slope of the surface from the point to the 

 River Somme at P, we are indeed surprised to find the sudden change 

 in gradients in passing southward from to Eenancourt, across the 

 now dry vaUey leading from Ferrieres (passing by Saveuse) to 

 Amiens, a distance of four or five miles. These escarpments con- 

 mence near Perrieres, and increase as the bottom of the valley 

 falls in a north-easterly direction towards the river Cette. I mea- 

 sured a section near the Perme de Grace, where the side of the dry 

 valley slopes at an angle of 20° to the bottom. 



These escarpments are better shown in a section taken from the 

 point 0, towards Eenancourt, and giving gradients from 30° to 50°, 

 representing flood-lines of former periods, but so sharp in defini- 

 tion that they look like the work of the last winter. (Fig. 7.) 



Fig. 7. — Section across the Saveuse valley. 



380 yards broad. 



28 yards deep. 



^ 



t^ ^ 



6' 







^■~~^--^^^. 







It is obvious that any theory of excavation of the Somme valley, 

 at Amiens, must take into account the, condition of the dry Saveuse 

 valley, which is only a type of hundreds of other dry valleys, 

 which formerly were filled with water falling into the Somme and 

 swelling it into a river capable of overflowing St. A.cheul. 



VOL. XXIV. PART I. K 



