492 ` The American Naturalist. [June, 
conformed exactly to the pre-existing ravine running in that 
direction, as this ravine would represent continuously the low- 
est point in the divide, hence we find the upper part of the 
present gorge showing no marked changes in its position. 
The ravine on the down slope toward the east, however, 
whose outline is represented on plate XIV by the dotted line,and 
on plate XV by the crossed line, fared quite differently. A large 
volume of water suddenly attempted to occupy a small ravine. 
The result was that the first curvature was greatly increased 
and a great undercut made at X. Deflected from this pointit 
struck the opposite or left bank of the ravine at m, and as it 
cut farther back into the great curve at OO, it also cut deeper 
and a small remnent of the left bank of the original ravine 
was left atm. The outline of this channel is represented by 
the broken line. 
After making the curve at OO where considerable under- 
cutting is also shown, the waters took a straight shoot south in- 
to the old ravine again. ane 
At the south end at YY, the old ravine, as shown by the 
dotted lines, made another sharp curve similar to the one at g, 
and passed north between S and T to another curve at H H, 
thus making the next loop of the sigmoid which passed south- 
east through the point u. 
Since the rush of glacial waters was not able to make the 
short turn at Y Y, and since the original surface level was 
lower than at H, they broke over the divide and made a new 
cross channel tending north-east in the direction Zu, an 
chocked up the old ravine between S and T with sediment. 
As the lower level was reached and the velocity checked, the 
lower courses of the intersected ravine were filled up with ma- 
terial cut from the gorge above. Evidence of these buried 
channels is found in many wells in the village of Toboso, just 
south of the river dam, which is across the present river ât F. 
There yet remains the great question, Why the river e 
left this second channel which it had cut out at such @ great 
depth and made a new one for itself straight across the gon 
barriers at gc’, HS, and KT. If these were at the ends oF 
the loops, or if the old channel was anywhere obstructed wi? 
