666 GEOLOGY OF OLD HAMPSHIRE COUNTY, MASS. 



and approach nearer to each other, and where they merge in "reticulate 

 ridges" this complexity becomes extreme. 



In a continuous section through a plain abounding in kettle-holes new- 

 ones came to light wliicli are not indicated by any depression on the sur- 

 face, they having been formed and filled before the completion of the 

 plain. In other cases this filling has been partial, and in every degree, so 

 that a very slight depression upon the surtace may indicate an extensive 

 disturbance of the beds below. 



In cases where the lamination of the sands is very regular and hori- 

 zontal the beds sink down on approaching a kettle-hole, with little disturb- 

 ance where the bending is slight; where it is considerable they dip inward 

 from all sides with a series of small faults ; where the subsidence is extreme 

 they are thrown into complete confusion; where they are wholly or partly 

 filled the upper layers thicken downward to fill the depression, and have 



Fio. 40.— Section soath of Millers Falls station to show kettle-holes formed by ice melting from beneath the sands. 



often a different structure in the depression from that outside, generally a 

 strong cross-bedding dipping in the direction of the current, which shows 

 that the sinking took place somewhat suddenly and the next flood found 

 and filled the depression rapidly. Where the whole thickness of the sand 

 was exposed I found it in one series resting upon a glacial surface of 

 gneiss and trap without the intervention of clay, which might have made 

 place for the sinking in of the sands b}-^ lateral flow, as has been suggested; 

 nor is there any indication of a lateral movement of the sands out from 

 under the sunken areas, as I have seen them undisturbed and horizontal on 

 all sides as the broad cutting removed the whole kettle-hole. 



Tliese depressions are found only in the lake bench or in the still 

 higher glacial lake beds, and are frequent in places, as in the center of broad 

 sand plains, where local eddies would be supposed to be least efi"ective. 

 Indeed, their great depth, great extent, very steep sides, and irregular distri- 

 bution would make it difiicult to explain them in this way. On the other 



