Maiv — On Sitbaerial and Marine Denudation. 441 



ample of marine excavation, the eroding current must have been per- 

 fectly constant in its exact location and direction during a range of coast 

 elevation or depression of at least 1200 feet, and you must assume 

 that there were five or six such independent currents within as many 

 miles to form the other similar indentations along the neighloouring- 

 coast. In connection with this locality another grave difficulty pre- 

 sents itself: you are not in the open sea, but the valleys opening inta 

 the coast debouch into the mouth of the Menai Straits, up and down 

 which a strong tidal current is constantly running transversely. The 

 Isle of Anglesea rises opposite within a few miles to a height of more 

 than three hundred feet, and must always have formed a barrier 

 against the sea directly assailing the point now occupied by the Aber 

 valley, to, at least, the extent of its present heigiit. Now, even if 

 the possibility is admitted of a current assailing the coast at Aber, 

 when Anglesea was beneath the sea, how are we to account for the 

 excavation of the valley to the extent of two or three hundred feet 

 below the level of the top of the opposite island -barrier ? These are 

 only a few of the many local instances in which the theory of 

 marine currents is easily shown to be inapplicable, but the whole 

 assumption that currents, whether of wind or water, can move up 

 and take effect at the extremity of a cul cle sac is fallacious ; you may 

 as consistently assume a power to make the smoke pass up a chimney 

 with the top closed ; there can be no motion loitliout a tliorouglifare. 

 This may be familiarly illustrated by the impotence of the wind to 

 enter an open window when the door of the room is closed, but 

 only make an outlet for the wind and its power through the room is 

 at once asserted. 



On the Diversion of Watershed Lines and Biver Channels. — In connec- 

 tion with the existence of transverse gorges, and the passage of river 

 channels through high land and mountain chains, it is important to 

 bear in mind the extreme antiquity of the general structure of the 

 present hill and valley system; on the first emergence of a sedi- 

 mentary deposit from the sea under which it was formed, the subaerial 

 waterflow begins to take up a definite position, and this may in the 

 first place be determined by the most trifling inequalities of surface. 

 The slightest elevation will throw off the rain, and river channels 

 will begin to be formed in the faintest depressions ; the water must 

 find an escape and it will leave the higher and find the lower ground, 

 however trifling the difference in height may be. We must thus 

 look for the initiation of the present watershed and waterflow sys- 

 tems ia the very earliest history of the emergence of the new-made 

 land, and when once formed they will maintain, under uniform cir- 

 cumstances, their original positions. 



The want of correspondence between river channels and the 

 general valley contour of a country, although not unfrequent, must 

 be looked upon as purely exceptional, and attributable to special cir- 

 cumstances that have supervened since the land surface received its 

 first impress of watershed and valley systems. Among the causes 

 tending to the diversion of river channels may be enumerated, frac- 

 ture or the opening of rents, drift accumulations, damming up 



