6/4 FRANK S. MILLS 



in its work and excavating in homogeneous drift, might also on 

 a return vibration undercut one or more terraces left by a previ- 

 ous shifting of its meanders. This would leave its vertical 

 downcutting, or erosion, expressed in a high escarpment, which 

 would represent the sum of later terrace scarps combined into a 

 single scarp coincident with one of an earlier terrace. 



Davis also calls attention to the fact that the preservation of 

 a series of terraces is due to the dominant influence and control- 

 ing power of buried ledges. These latter have the character of 

 an upland spur projecting transversely out into the valley. A 

 degrading stream in alluvial deposits, finding itself superimposed 

 upon a transverse rock spur, would tend to migrate farther 

 toward the axis of the valley, and thus slip by the obstruction, 

 provided it was not so far incised laterally as to be held fast by 

 the spur. The succeeding meanders, degrading having gone on, 

 would return at a lower level and impinge upon the intercepting 

 ledge still nearer the axis of the valley at each successive swing. 

 On the other hand, a down-sweeping meander having encountered 

 the buried ledge, and having a curvature above it well developed 

 through deep lateral cutting, might be held fast by the projecting 

 rock spur and unable to slip by. Succeeding meanders having 

 advanced down stream, and their curvature above the ledge hav- 

 ing been highly compressed, they would undercut and scour out 

 former terracing by the enforced increase in the radius of its 

 lateral cutting. This might continue until the meander became 

 so compressed that it would eventually destroy itself by a cut- 

 off, and allow the stream to pass the ledge by straightening its 

 course. The presence of such spurs and buried ledges would 

 defend and preserve on its immediate lee, or down-valley side, 

 any terracing cut at an earlier period in the drift covering above 

 it. Thus a salient of terraces would be preserved above the 

 rock defended cusps. 



The Catatonk River terraces do not lend themselves to such 

 interpretations, either as being the work of a stream of constant 

 volume, or their preservation being due to the intervention of a 

 protecting spur (Fig. 3). 



If we agree that the radius of a meander curve is limited and 



