AMMux op sciences] NIAGARA AND THE GREAT LAKES 221 



this phase of the problem as it lay too far south of the Great Lakes. On the other hand, where 

 the land sloped toward the ice sheet it would form a barrier and obstruct the drainage, with the 

 result of producing proglacial lakes, such as have already been considered in connection with the 

 shore lines on which Gilbert worked so long. The outlets of such lakes would be either over 

 the lowest pass in the hills on their southern border — and such was the case with the proglacial 

 lake of the Maumee valley, the outlet of which ran southwestward across Ohio to the Mississippi 

 system — or along the depression that must have existed between the iceward slope of the land and 

 the landward slope of the ice at one side or the other of the lake. In the second place, it must 

 be understood that the Appalachian plateau in west-central New York is built up of essentially 

 horizontal shales and limestones, and that its north-sloping margin is divided into broad spurs 

 by wide-spaced valleys. When the retreating ice sheet lay against the plateau slope, a narrow 

 lake must have occupied each of the valleys, and the lake outlet would have been either south- 

 ward over a pass at the valley head to the Susquehanna system or laterally along the ice-margin 

 depression to a lower lake on the west or east. 



Let it be supposed that the lakes here to be considered all have lateral discharge. One of 

 them must have been the highest of the series. Whichever way it discharged, all the little lakes 

 to the west of it must have discharged westward to lakes of lower and lower level, and all those 

 to the east of it must have similarly discharged eastward. While the lakes were of small area, 

 their outflowing stream along the ice margin would not have eroded large channels, or "doughs" 

 as they might have been called, across the plateau spurs; but a time came when, by reason of 

 the ice retreat, all the lakes to the west of the highest point in the ice-margin depression were 

 united in one, and when that one vast water body nowhere had a westward or southward 

 discharge so low as the eastward discharge at that highest point in the depression. Then and 

 thereafter the channels would have been eroded across the plateau spurs by a great outflowing 

 river as it ran to lower and lower east-discharging lakes, and the channels would therefore 

 have gained impressive dimensions. This is of course all highly imaginary, and only when cer- 

 tain minutely specialized elements of so imaginary a scheme are found to be verified by the 

 occurrence of then counterparts in nature can its verity be accepted. Hence it was to those 

 specialized elements that Gilbert gave particular attention. In doing so he showed exceptional 

 keenness of penetration and an altogether unusual capacity of visualizing a moving picture of the 

 past. Some of the specialized elements may be described. 



If a channel is cut eastward across a spur from a higher to a lower ice-barred lake, the 

 west end of the channel ought now, as Gilbert phrased it, to begin "in the air," to fall from west 

 to east, and to have a delta where its east end opens into a north-south valley. Furthermore, 

 the channel ought to exhibit features characteristic of channels now occupied by rivers, such as 

 steeper slopes over its concave banks, falls where hard strata are cut through into weak strata, 

 and a phinge pool in the weak strata at the base of the falls. Again, the delta at the eastern or 

 discharging end of a channel in one spur ought to stand close to the level of the west-end intake of 

 the corresponding channel in the next spur to the east. All of these deduced elements of the 

 scheme were repeatedly verified ; and in some instances the verifying features were found to be 

 tremendously impressive. One of the channel-cutting streams which crossed a broad plateau 

 spur south of Syracuse appears to have taken a consequent course that locally led it to make a 

 south-turning loop around a drumlin which was superposed on the plateau slopes; the present 

 form of the channel shows that the stream proceeded to cut laterally into and thus to steepen the 

 concave banks on the outside of its several curves, thereby enlarging the loop and at the same 

 time narrowing its neck, until the stream eventually opened a out-off passage through the 

 narrowed neck, just as normal rivers do when they cut off a knob within a turn of an incised 

 meandering valley. In another channel an especially fine cataract must have been developed: 

 an upper-level channel bed of bare limestone leads to the top of the cataract cliffs, the cliffs 

 descend into the amphitheatral head of a lower-level canyon, the plunge pool below the cliffs is 

 now occupied by a little lake, at the farther end of the canyon a delta flat is built out into an 

 open north-south plateau valley, and at the level of the delta but on the farther side of the valley 

 lies the intake of another channel in the next plateau spur. All this is so clear that one can easily 



