OLD MIDDLE ALLEGHENY DRAINAGE SYSTEM. 137 



been opened to its present depth and had southward drainage before the 

 beginning of the glacial period. These gravels are in every observed 

 case situated on sloping points on the inner curves of sharp bends in the 

 river. At such places a stream works outward as well as downward, there 

 being erosion on the outer curve and liability of deposition on the inner 

 curve. It is to be expected, therefore, on the hypothesis that the stream 

 has greatly deepened its channel since the ice invasion, that such deposits 

 should be present, and these deposits do not, it is thought, necessarily 

 oppose the hypothesis of former northwestward drainage, nor that of great 

 erosion since the beginning of the Glacial epoch. 



Concerning the line of discharge for the Middle Allegheny from near 

 Meadville to the Lake Erie Basin a few remarks seem necessary. It is 

 certain that the old drainage line did not follow French Creek Valley 

 northward beyond Meadville, for there is clear evidence of an old divide 

 on the present creek a short distance above that city. The line described 

 by Carll as the Coimeaut outlet departed from French Creek about 4 miles 

 below Meadville, followed up the outlet of Conneaut Lake to that body of 

 water, passed northward across a low divide filled heavily with drift to the 

 northward-flowing portion of Conneaut Creek, passed down that creek to 

 the bend near Albion, then continued northward and entered the Lake Erie 

 Basin a few miles east of the Ohio-Pennsylvania line. 



Another valley-like lowland leads from Meadville along Cussewago 

 Creek (reversed) nearly to its source, and thence northwestward to the 

 Conneaut outlet near Albion, through a region heavily covered with drift. 

 The Cussewago channel is narrower than the Conneaut and seems, on the 

 whole, a less probable line of discharge for the Middle Allegheny. The 

 northern end, however, aff'orded a line of discharge for a jDortion of the 

 French Creek valley above Meadville, as indicated on page 139. 



On both the Conneaut and Cussewago channels the borings are too 

 few to afiFord a satisfactory knowledge of the rock floor. At the border of 

 Lake Erie, for several miles each side of the place where the old stream 

 entered, the rock surface seldom rises above lake level. It is probable 

 that the channel of the old stream had reached a level in harmony with 

 the bed of the lake. 



It was suggested by CarlP that the headwater portion of the Shenango 



' Second Geol. Survey Pennsylvania, Kept. I^, pp. 5-6. 



