358 J. G. Goodchild — On Glacial Erosion. 



along the line ; and each must also have accomplished only a certain 

 definite amount of cutting back before its position again changed. 

 Even granting that it is possible for springs to cut back an escarp- 

 ment with a certain total amount of regularity, it is clear that, unless 

 there are at hand larger streams to remove the undermined rock, the 

 springs would soon become so much choked up that their action 

 would no longer be effectual. In other words, if the rate of removal 

 do not keep pace with the rate of disintegration, the accumulating 

 talus would soon protect that part of the scar from much further 

 alteration. Another objection to the spring theory is that nearly all 

 the denudation effected by the springs would be confined to beds 

 above the impervious bed that throws them out ; hence, if the lime- 

 stone scars are really the result of spring action, we ought to find the 

 limestone scar in all eases at a considerable distance nearer the 

 centre of the hill than the outcropping edge of the sandstone that 

 forms its base. Instead of this, we find, in nearly every case but 

 that where the limestone directly overlies a soft bed, that the lime- 

 stone scar forms part of one continuous slope with the outcrop of the 

 sandstone beneath. Hence the conclusion seems inevitable that both 

 scars were formed at the same time and by the same agencies. 

 Lastly, some of the terraces whose origin is here discussed have a 

 width of two hundred or three hundred yards, or even more than 

 that ; yet the outer part of the terrace, that is to say, the part nearest 

 the marginal scar, usually exhibits no greater amount of weathering 

 than does the innermost part close to where the next bed above 

 comes on. If then the terraces have really been produced by Sub- 

 aerial Erosion acting unequally upon beds of different degrees of 

 destructibility, it follows either that the overlying shales have been 

 cut back from the limestone at a greater rate than we have evidence 

 for ; or else, if the time occupied in cutting them back has been long, 

 the weather has no appreciable effect upon the limestone after this 

 has undergone a certain amount of subaerial erosion. Yet there is 

 clear proof that, under circumstances the most favourable, so soft a 

 rock as shale has rarely been cut back far in Post- Glacial times ; 

 therefore it is clearly impossible that a much larger quantity can 

 have been removed in the same time in situations where no stream 

 could possibly flow under anything like the present physical con- 

 ditions. The alternative that after a time the limestone ceases to 

 undergo any further erosion needs no other argument to disprove it 

 than that afforded by the position and shape of the swallow-holes. 

 Under subaerial conditions the streams that gave rise to the swallow- 

 holes must be continually cutting back the soft beds overlying 

 the limestone from the point where the stream first reached the 

 limestone towards the watershed. Consequently, the swallow-holes 

 that were first formed would either remain in their original shape 

 and position to mark where the stream first began to sink, or else 

 would tend to lengthen in the direction of the source of the stream, 

 as the point where this first reached the limestone slowly receded 

 towards the watershed. Hence the swallow-holes, instead of 

 retaining a rudely circular form, would change first into the form 



