222 GROVE KARL GILBERT— DAVIS [MemoirS [vo A l T xxi; 



imagine the rushing river still at work, sweeping along the upper level channel, plunging down 

 where the channel is cut off by the cliffs, whirling and foaming in the pool below, and then dashing 

 away through the low-level canyon to the delta in the long and narrow lake of the next north- 

 south valley. 



RELATIONS OF SUCCESSIVE CROSS-SPUR CHANNELS 



But this is not all. The ice margin was not fixed ; it was slowly retreating northward down 

 the plateau slope. Hence in time a new ice-margin path across a spur might be offered, not only 

 at a lower level than that of the first path adopted by the east-flowing lake outlet, but at a lower 

 level than that of the channel floor which the outlet river was cutting down in the spur while 

 the ice was retreating; for be it understood that when the river first ran across the transverse 

 profile of the spur the channel was so short that it could be deepened rapidly, but the deeper it 

 was cut, the longer it would become, and the slower it would be worn down. Hence rapid channel 

 deepening might at first hold the outlet river for a time on a course that was deepened faster than 

 the path along the retreating ice margin was lowered; but eventually the continued northward 

 lowering of the ice-margin path would enable it to overtake the retarded deepening of the chan- 

 nel ; thereupon the outlet river would desert the first-cut channel and begin cutting a second one 

 farther down the spur slope. This process might be repeated several times, so that a single spur 

 would be transversely trenched across by a number of separate channels. Now a definite 

 relation as to altitude should obtain among the channels thus carved ; namely, the point on the 

 sloping spur crest where a new channel is adopted next to the ice margin must be at least as low 

 as the west-end intake of the abandoned channel farther up the spur slope. This definite relation 

 is so highly specialized that, if the facts confirm it, the theory from which it is deduced will 

 demand acceptance. The facts do confirm it, and not only once, but over and over again, and 

 Gilbert's explanation of the channels thus becomes irresistible; but in some instances the 

 occurrence of till in a channel indicates that the retreating ice margin advanced again a short 

 distance in its retreat. With the adoption of the explanation the glacial theory as a whole gained 

 yet another group of facts for its support; facts whpse existence had been individually and 

 locally known for years, but whose real relations were never understood until Gilbert revealed 

 them. 



A LONG CHANNEL FLOOR AS A GREAT HIGHWAY 



The cross-spur channels are in general rather small affairs. They are too small to be 

 represented on ordinary State maps, although they appear clearly enough on the inch-to-a-mile 

 topographic maps later issued by the national survey; but at the time of Gilbert's studies he 

 had no such maps of his area and was much hampered by the lack of them. He expressed the 

 need of them in a letter to a New York friend : 



If you have the ear of a State legislator, fill it full of the importance of topographic surveys — we all want 

 those contour sheets. 



That was about as near as Gilbert ever came to lobbying. Now that the maps are pub- 

 lished, it is easy to see that certain cross-spur channels have a local significance when serving 

 as low-grade roadways through the plateau spurs; and a few of them are advantageously used 

 by railways as natural cuts. One of the most remarkable examples of an east-west channel 

 thus utilized runs for nearly 30 miles, not across the plateau spurs, but through a part of the 

 drumlin belt that occupies the lower land north of the plateau margin between Syracuse and 

 Rochester; Gilbert had discovered this long channel a year earlier than the more strongly marked 

 channels in the plateau spurs, and then wrote to a friend of the stretch from Palmyra past 

 Lyons to Clyde as follows: 



You will find yourself in a valley whose dimensions are not determined in the least by the small stream 

 flowing through it. It is of tolerably uniform cross-section, with steep walls, and it traverses the country with large 

 sweeping curves. It is not, to any considerable extent at least, bounded by rock, but the walls are of drift, 

 chiefly drumlins; and I think you cannot follow it far without being satisfied that the drumlins have been cut 

 off by the agency producing the valley. This agency I understand to be the flow of a great body of water, a 

 body large enough to determine by its momentum that the changes of direction should be in curves of large 

 radius. 



