96 GLACIAL FORMATIONS OF ERIE AND OHIO BASINS. 



similar constriction at the very mouth of the river, just below Newcastle. 

 Hills 300 to 400 feet in height are separated by a channel scarcely one- 

 third of a mile wide at the level of the river, which is not far from the level 

 of the gradation plain. Between these two valleys there is an abandoned 

 channel leading north from the Mahoning at Edenburg, Pa., to the 

 Shenango at Harbor Bridge, 4 miles above Newcastle, and this is a httle 

 wider than either of the other channels, though oidy about half as wide 

 as the gradation plain on the Beaver. The continuation of the channel 

 northward up the Shenango is in a valley one-half to three-fourths as wide 

 as that of the Beaver gradation plain. From the Shenango at Sharon, Pa., 

 there is an abandoned channel running westward to the Mahoning at 

 Youngstown, Ohio, from which point a broad valley opens northwestward 

 into the Grand River Basin, and this in turn opens into the Lake Erie Basin. 



Since the constricted portions of the Mahoning and Shenango valleys 

 may appear to oppose the hypothesis of a discharge of the old Monongahela 

 system through them, they will be considered before the routes just out- 

 lined are discussed. 



From descriptions given by White ^ it appears that the rocks forming 

 the Carboniferous conglomerate measures are exceptionally soft for a few 

 miles along the Beaver Valley, below the junction of the Mahoning and 

 the Shenango, where the gradation plain is broadest, but contain firm and 

 resistant beds of considerable thickness in the constricted part of the 

 Shenango in the vicinity of Newcastle, and also on the Mahoning at its 

 constricted part near the State line, l^hese beds are also very firm and 

 resistant on the Beaver in the vicinity of Homewood, but they dip rapidly 

 southward and soon pass below the level of the old gradation plain. They 

 cause a notable constriction in the inner valley or trench south of Home- 

 wood, but lie mainly below the level of the broad gradation plain. The 

 occurrence of ledges which are more firm and resistant in the narrow than 

 in the broad portions of the old gradation plain tends to greatly reduce, if 

 not remove entirely, the difficulty of carrying through a line of northward 

 drainage. Concerning the variability of the upper part of this rock series 

 White makes the following remai-ks: ^ 



The rock in question is most variable. Only 3 miles south from the Lawrence 

 County line we find it 155 feet thick in the great ledge at Homewood, but in coming 



^Second Geol. Survey Pennsylvania, Rept. Q^ 1879. ^Ibid., pp. 53-54. 



