276 THE WONDERS OF GEOLOGY. Lect. III. 



the beds of lava and limestone to the primary rocks beneath, 

 and intersect the country with valleys and ravines, spreading 

 over the ancient beds a thick covering of alluvial soil.* 



53. Excavation of valleys by streams and rivers. 

 — There is no district which exhibits in more striking cha- 

 racters the erosive power of running water, than Auvergne. 

 In many places the basalt is columnar, like that of Staffa, 

 and the Giants' Causeway ; and one range, on the banks 

 of the Ardeche, forms a majestic colonnade 150 feet in 

 height, extending a mile and a half along the valley which 

 has been channelled out by the river that flows at its base. 



Mr. Scrope's description of this process is highly gra- 

 phic. " The bed of the Ardeche is strewed with basaltic 

 boulders, pebbles, and sand, originating from the destruc- 

 tion of the columnar ranges. In some of the volcanic cones 

 the beds of basalt may be traced issuing from the crater 

 and following the inequalities of the valley, just as a 

 stream of lava would flow down the same course at the 

 present time. Yet these ancient currents have subsequently 

 been corroded by rivers which have worn through a mass 

 of rock 150 feet in height, and formed a channel even in 

 the gneiss rocks beneath, since the lava first flowed into 

 the valley. In another spot, a bed of basalt, 160 feet high, 

 has been cut through by a mountain stream, and very 

 beautiful columnar masses are exposed. The vast excava- 

 tions effected by the erosive power of water along the 

 valleys which feed the Ardeche, since their invasion by 

 lava currents, prove that even the most recent of these 

 volcanic eruptions belong to an era incalculably remote." 



54, Extinct volcanoes of the Rhine. — I have dwelt 



* This account of the volcanic phenomena of Auvergne, is an abstract 

 of the interesting Essays of Messrs. Bakewell, Scrope, Lyell, Mur- 

 chison, Dr. Daubeny, Dr. Hibbert, MM. Croiset, Jobert, Robert, and 

 Bertrand-Roux. Mr. Scrope's work cannot be perused, even by the 

 general reader, without deep interest. 



