KETTLE-HOLES. 667 



hanil, all tlu-ir pcvuliarities seem to me to find al)un(lant explanation in the 

 assumption that reuniants of the ice became covered by the sands, and in 

 the then low temperature remained often for a long time, as in the arctic 

 countries to-day; and as my knowledge of the region has become more 

 extensive I have found more and more evidence of the presence of ice still 

 in the valley during the building of the high terrace. 



The accompanying figures show in detail the facts summarized in the 

 preceding paragraphs. They are selected from a much larger number, and 

 were in part drawn with the thermometer much below zero and are given 

 just as they were made. 



In fig. 40, showing the long section on the New London Northern 

 Railroad extending south from the station at Miller.s Falls and taken from 

 the east wall of the cutting, the upper layer is a coarse gravel 5 to 7 feet 

 thick, coarsest and thickest at the north end of the section, where it is nearest 

 the source of supply at the head of the delta. Beneath this are fine white 

 sands which run in thick sheets horizontally for long distances, the sheets 

 showing the finest false bedding. It is perfectly plain that the gravels were 

 laid down on a horizontal surface of the fine sands, and that these were 

 thrown down in horizontal sheets, and that the sink-holes are of later origin. 



At 1 in the figure is the beginning of the cutting just below the station. 

 It is the southern border of an irregular sunken area that sends lobes east 

 nearlv to the rocks. The fine sands sink gradually below the surface, and 

 the great thickening of the gravels may be due in whole or part to a sinking 

 of the sands before the whole of the gravels were tlu'own down. 



At 2 the sinking of the under sands is accompanied by a series of 

 .small parallel faults, dij)ping inwardly from the irregular sinking, and the 

 gravel thickens downward, parti}" filling the depression. At 3 is a wholly 

 submerged sink-hole, the sands bending down and the gravels thickening 

 down to fill the space, showing that the ice melted away before the deposi- 

 tion of the upper beds. At 4 and 5 this is repeated. At 6 is a fine kettle- 

 hole, and marked faults dipping inward from both sides, accompanied by 

 a great number of smaller faults with throw in the same direction. At 7 

 the faults are complicated, as if by gradual melting a new substratum had 

 several times been produced during the building up of the sands. 



In the lower figure is shown the other side of the cutting, so that the 

 two views face each other, as shown by the cross section. Here the kettle- 



