278 Chamberlin and Leverett — Studies of the 



are really residues of a re-filled valley, they would seem to 

 point to such an interpretation. 



Low Rock Shelves. — Along the lower Allegheny, there 

 stand a number of rock shelves at about 75-100 feet above the 

 present river. These are of some considerable extent, occa- 

 sionally reaching to a quarter of a mile in breadth. While 

 not all at a common horizon, they are sufficiently near to be 

 fairly correlated. There is but little drift on these. They 

 may mean merely an incident in the cutting down of the val- 

 ley, or they may signify the termination of an epoch of 

 degradation, followed by a notable halt or possibly a re-filling. 

 We have found little to suggest whether or not they have 

 much significance and so pass them for the present with sim- 

 ple mention. 



Interpretations. 



The foregoing outline falls far short of properly setting 

 forth the facts, and yet our time is far spent. We neverthe- 

 less beg a few words respecting interpretations. There is not 

 time, however, for the discussion of these, and we shall con- 

 tent ourselves with a simple statement of such hypotheses as 

 seem to have any justification in the phenomena now known, 

 and with some comparisons respecting their import. 



To us it seems clear that all hypotheses that do not explain 

 the phenomena by simple glacial and fluvial action (with inci- 

 dental ponding), are excluded by the very characteristics of 

 the deposits themselves. We have reason to believe that this 

 view will be no longer seriously contested. 



We are also clear in the conviction that no hypothesis of 

 continuity can explain the phenomena. They seem to demand 

 unequivocally important stages of deposition separated by im- 

 portant stages of excavation. 



Of hypotheses which fall within these limits, there are three 

 which postulate a single pair of depositions and of erosions, 

 and a fourth which postulates a triple alternation. The first 

 group fall into accord with the twofold division of the adja- 

 cent drift, the attenuated drift, and the moraine-bordered 

 drift. The third conforms to the threefold division of the 

 drift, which is found in western Ohio and beyond. For con- 

 venience we shall state these hypotheses as applied to the 

 more regular part of the system below the mouth of the 

 Clarion. The qualifications for the more irregular upper por- 

 tion may be readily derived from what we have said respect- 

 ing those portions. 



Hypothesis I. — This hypothesis is based upon a minimum 

 expenditure of dynamic action, though this is an after thought 

 and not the parent -of the hypothesis. By this we mean that 



