276 Chamberlin and Leverett — Studies of the 



on the convexities of its bed and is disposed to permit the 

 lodgment of material on its concave side where the tendency 

 of the stream is to recede. Now, the Allegheny during the 

 whole progress of its descent from the level of the high ter- 

 races to its present position, was undoubtedly a gravel-bearing 

 stream. It was not only engaged in the process of removal of 

 the gravel along its own immediate course, but was receiving 

 very much that was washed in from the drift region adjacent, 

 so that a certain amount of lodgment of transported material 

 may be assumed to have been inevitable. This is dependent 

 upon the same principle of action that would permit the re- 

 tention of gravels, in such situations, if they had been pre- 

 viously deposited within the trench. As the gravels on the 

 slopes are usually thin sheets or patches, we have not found 

 decisive evidence, in themselves, as to whether they are rem- 

 nants of earlier gravels, or incidents of degradation. We have 

 searched industriously for evidence that should be decisive on 

 this point. Such evidence should be found in abandoned seg- 

 ments of the old valley, if it had been deeply excavated before 

 the deposit of the gravels and had subsequently been filled by 

 these up to the summit of high gravels. These high gravels 

 fill ox-bows and recessed shelves, and the stream which depos- 

 ited them had, in many instances, alternative courses. This is 

 notably true in the vicinity of Pittsburg. Here the old high 

 plane of rock was extensively covered by the waters that de- 

 posited the gravels, as is shown by the presence of remnants. 

 There are, in the eastern part of the city, four islands sur- 

 rounded by broad channel-ways, among which the waters dis- 

 tributed glacial gravels in greater or less degree. Now, if the 

 present deep Allegheny and Monongahela trenches had been 

 cut previously to the filling in of the gravels, there is only a 

 small chance that, after the gravel- depositing period, during 

 which they were flowing 50 feet or more above the rock plane, 

 they would have descended the second time on precisely the 

 same lines. Between the several broad channels open to them, 

 the possible combinations are 32 in number, and hence theo- 

 retically the chances of a combination repeating itself are one 

 in 32. If it be objected that certain of the courses are more 

 favorably situated than others, our answer is, first, why were 

 these others then ever produced by the streams or occupied by 

 glacial wash, and our second answer is, that, if this be true of 

 certain combinations, it does not seem to us to be at all true of 

 many others. 



Besides this, along the Allegheny river above, and also 

 along the Ohio river between Pittsburg and Toronto, to which 

 point the high gravels containing Canadian pebbles have been 

 traced, there are perhaps a score of ox-bows, deep recesses, 



