OF THE RIVER-DRIFT GRAVELS. 291 



granitic debris brought down by the Yonne forms a notable 

 proportion of the gravel ; and at Precy, near Creil on the 

 Oise, the fragments of the ancient rocks are abundant ; 

 but lower down the Seine at Mantes, they diminish very 

 much in quantity, and at Rouen and Pont de 1'Arche I 

 saw none, though a longer search would doubtless have shown 

 fragments of them. This case of the Oise is however in- 

 teresting, not only on account of the valuable evidence con- 

 tained in the above quotation ; but because, though the river 

 flows, as a glance at the map will show, immediately across 

 and at right angles to the Somme, yet none of the ancient 

 rocks which form the valley of the Oise have supplied any 

 debris to the valley of the Somme : and this, though the two 

 rivers are at one point within six miles of one another, and 

 separated by a ridge only eighty feet in height. 



The same division occurs between the Seine and the Loire ; 

 " Bien que la ligne de partage des eaux de la Loire et de la 

 Seine, entre St. Amand (Nievre) et Artenay, au nord d'Orleans, 

 soit a peine sensible, aucun debris de roches venant du centre 

 de la France, par la vallee de la Loire n'est passe dans le 

 bassin de la Seine.* 



In the Vivarais near Auvergne, "Les depots diluviens 

 sont composes des memes roches que celles que les rivieres 

 actuelles entrainent dans les vallees, et sont les debris des- 

 seules montagnes de la Lozin, du Tanargue et du Mezene, 

 qui entourent le bassin du Vivarais." f 



Again, ^ 



" Le diluvium des vallees de 1'Aisne et de 1'Aire ne ren- 

 ferme que les debris plus ou moins roules des terrains que 

 ces rivieres coupent dans leur cours." J 



Finally, Mr. Prestwich has pointed out that the same 

 thing holds good in various English rivers. The conclusion 



* D'Archiac, I.e. p. 164. f D'Archiac, I.e. p. 160. 



J Jtfalbos. Bull. Geol. Yol. iii. p. 631. 



