266 Chamberlin and Leverett — Studies of the 



oped along Buffalo creek and also north of its mouth. At 

 Beech Bottom, West Virginia (three miles below Wellsburg), 

 is a broad shelf at 935 feet ; at Portland, Ohio (four miles far- 

 ther), there are two notable shelves, the higher at 1085 feet 

 and the lower at 950 feet; opposite Wheeling, West Virginia 

 (eight miles below Portland), there is a notable terrace at 930 

 feet and another at 990 feet, while back of Wheeling a plain 

 sweeps about the high hills at about 1100 feet. 



Here there seems irregularity and want of system, but if the 930 

 feet terrace at Wheeling, the 950 feet terrace at Portland, and the 

 935 feet terrace at Beech Bottom be connected with the 950 feet 

 terrace at Toronto where the glacial gravels cease, and these be 

 interpreted as remnants of the trench cut across the old divide 

 after the supposed reversal and while the gravels above were 

 being accumulated, and if all the rest be interpreted as rem- 

 nants of the floor of the old system running northward to join 

 the lower Allegheny-Monongahela-Beaver system, the whole 

 seems to fall into tolerable harmony. The terraces referred to 

 the trench below Toronto are generally swept clean of debris 

 and so accord with the hypothesis. The breadth of the cor- 

 responding gradation-plane at the mouths of tributaries seems 

 larger than might be expected under this hypothesis unless we 

 make the. first glacial epoch long, and this is, perhaps, against 

 the hypothesis. We leave the question open for the present. 



Dr. Foshay also called attention to the low altitude of the 

 present rock floor of the valley along the supposed outlet 

 (there being oil well records showing lower rock bottom on 

 the Mahoning than at the mouth of the Beaver), and he main- 

 tained that northward discharge was continued until the 

 streams had reached this low level. In opposition to this 

 view, it has since been determined by Mr. R. R. Hice that the 

 rock floor does not have a regular descent northward from the 

 mouth of the Beaver, but stands 60 to 70 feet higher in the 

 gorge near the mouth of the Connoquennessing than at the 

 mouth of the Beaver, and about 90 feet higher than at Eden- 

 burg, about 15 miles up stream. His data seem complete, as 

 the piers of a railway bridge were built upon the rock floor 

 and so distributed as to test the middle as well as borders of 

 the valley. So far as collected, the data from well sections, 

 both on the Mahoning and Shenango, oppose the view that the 

 low rock floor near Edenburg was occupied by a stream which 

 discharged into the Lake Erie basin (see sections A and B, 

 fig. 3). The Mahoning, just above the very deeply excavated 

 portion, has a narrow valley 80 to 100 rods wide, or scarcely 

 one-half of its width where deeply excavated. From the bor- 

 ders of this narrow portion, tributary streams descend over 

 cascades from levels of 100 to 200 feet above the river. A 



