684 GEOLOGY OF OLD HAMPSHIRE COUNTY, MASS. 



lu the frozen wall the sand layers were deeply worked ont by the 

 wind, and the clay layers projected one above another like the eaves of a 

 fluted iron roof. By the thinning of the sand layers the whole deposit 

 loses in thickness and the superincumbent clay sinks lower. This plainly 

 indicates a strong current in the summer floods, a quiet one during the 

 winter, and a depth of water sufficient to so far remove the banks of the 

 stream to the west that the floods brought only thin layers of sand out over 

 the clay to this point, layers which have mostly dwindled to nothing before 

 we reach the south end of the cutting. 



The sands, as indicated above, nin up on the south slope of the drumlin 

 in a thin film, and could in one section be traced almost or quite continu- 

 ously across it, to join a thick bed of similar sand, which extends to the 

 north end of the section, where it is cut off by erosion. It is capped, as' in 

 the layer south of the drumlin, by the same thick deposit of clay. These 

 sands are finer than those already described, especially near the drumlin, 

 manifestly because they were laid down in the sheltered ai'ea behind it. 

 Northward they grow coarser, and at the extreme north are gravelly, and 

 iron-shot ^^■here the water circulated below the clays, with coarse cross- 

 bedding dljjping south. Through most of the distance the beds are 

 (or were) horizontal, and show repeated oscillations of coarser and finer 

 layers, and eveiy where most delicate cross-bedding. Upward, the whole 

 gradually becomes finer, clay layers making their appearance, which at the 

 end effect a somewhat sudden transition into the clay above; in short, the 

 sands agree in all points with the corresponding sands south of the drumlin. 



The fJiird fill. — It will make clearer the complex series we are studying 

 if I call attention to the four surfaces on which the ice has rested. The first 

 is the surface of the drumlin. The second is the surface of the second till, 

 which has eroded the pink sands ; and as the till layer seems largely derived 

 from the broad surface of the di-umlin, this layer lessens and the ice rests 

 almost directly on the sands in the continuation of the surface southward. 

 Consideration of the third till, which is the subject of discussion here, may 

 be omitted for a moment. The fourth and last surface occupied by the ice is 

 very clearly defined ah mg the whole length of the section. It is the hori- 

 zontal upper surface of the clays above the second sands from the north end 

 of the section to the drumlin, and is continued along the eroded surface of 

 the drumlin, and is the surface of the fourth till from the beginning of the 



