LYELL. 317 



choked up for a long time. A lake was formed, 

 and the river wore a passage between the lava and 

 the granitic schist, but the former was so exces- 

 sively compact, that the schist evidently suffered 

 most. In the progress of ages, the igneous rock, 

 one hundred and fifty feet deep, was cut through, 

 and the river went on and ate its way, thirty-five, 

 forty-five, and in one place eighty-five feet into the 

 subjacent granitic beds, leaving on one bank a 

 perpendicular wall of basaltic lava towering over 

 the gneiss. In the Vivarrais, where similar phe- 

 nomena had been observed, Herschel had remarked 

 a bed of pebbles between the lava and the gneiss, 

 marking the ancient river-bed, but Buckland en- 

 deavoured to get over this difficulty by saying that 

 these pebbles might have covered a sloping bank 

 when the river filled the valley, and that this bank 

 may have always been high above the river bed ; 

 for if the sloping sides of a valley, said the 

 Professor, be covered with pebbles, as they often 

 are, and the valley is filled with lava, and then the 

 lava cut through and partially removed, there will 

 of course be a line of pebbles at the junction of the 

 lava and the rock beneath, but these pebbles will 

 not mark an ancient river bed. Now, unluckily 

 for the doctor in this case, he has no loophole ; an 

 old lead mine, said to have been worked by the 

 Romans, happens to have exactly laid open the line 

 of contact, and the pebble bed of the old river is seen 

 going in under the lava, horizontally, for nearly fifty 

 feet. This is an astonishing proof of what a river 



