PHYSICAL C;EOGRAriIY OF THE KECilOX. 83 



encountered, it cftectively checks the deepening of the channel in the loose 

 detritus fin-thor up strciun, hut hclow it, the soft valley-filling is quickly 

 cut away as low as the next down-stream ledge will allow ; at the end of 

 every level stretch thus formed, there is a sudden descent over a harrier 

 to another level stretch below, and thus are produced the alternate flood- 

 plain meadows and low rocky rapids that characterize our rivers. Some- 

 times a river, wandering too far from its old line, unwittingly sinks its 

 channel on a spur high up on the buried valley-slopes ; then a cascade of 

 strong fall is formed, of vast importance to New England in furnishing 

 a\ailal)lc water-power, as at Manchester on the Merrimack, and manv 

 other similar points. If the valley be clogged with till, the stream will cut 

 a gradual descent through it, rushing down impetuously among the bowl- 

 ders that remain in its bed : thus the Contoocook flows below Hillsboro' 

 Bridge, N. H., and even the Connecticut has stretches of rocky and stony 

 channel through the Fifteen-mile Falls above Newbiuy. When rock in 

 place is disclosed beneath or near the till, pot-holes are often worn in it by 

 the bowlders, as at the upper falls of the Ammonoosuc, above Fabyan's, 

 or at Shelburne Falls on the Deerfield. 



Streams of intermediate size, on the more open, lower country southeast 

 of the mountains, have their courses so greatly influenced by glacial deposits 

 that they cannot be regarded as the successors of any corresponding pre- 

 glacial streams ; they flow irregularly among the rocky hills, drumlins, 

 kames and sand-plains that were disclosed for their settlement as the ice 

 wasted away, here meandering about a flat meadow that conceals some old 

 channel, there crossing over an old rocky spur or divide, or cutting down 

 a stony dam of bowlder-clay, but nowhere presenting that evident relation 

 between stream-volume and valley-size that prevails so manifestly in regions 

 of a simpler history, like West Virginia. Except the terracing in the 

 meadows and the slight gorge-cvitting on the old divides and spurs, the 

 surface drained by these new streams has not been developed under their 

 guidance ; it was presented to them ready made, and they are just making 

 tlieir first mark upon it. Oliverian Brook is a small example of a stream 

 thus thrown over an old divide : it descends southward from Aloosilauke, 

 as if to join a branch of the Pemigewasset, but abruptly turns northwest- 

 ward across an ancient pass, cutting a little gorge in the rock, where everv 

 traxeller in the Montreal railway may see it, and then flowing through 

 meadows to the Connecticut by Haverhill. The gorges of this class are 

 sludlower than their relatives in Xew York Ijccause the rocks here are 

 harder. The "flumes" of the White Mountains are not gorges of this 

 oriijin, but are cut out on vertical dikes that are weaker than the enclosin<j 

 country rock. 



The Saco may serve as a larger type of a new river, except in its up[)er 

 course where it follows a deep old valley out of the mountains to Conway ; 



