INFERENCES. 49 



northern part, and perhaps some limited spots where the village stands, has once 

 constituted a part of the channel of the stream. 



In the extreme northern part of Deerfield, only a mile south of the village of 

 Greenfield, occurs an old rocky bed of Green river, a tributary of Deerfield river. 

 Here are pot-holes in the red sandstone, and a gorge in the same, Avhile the pre- 

 sent river runs in a channel worn in sand and clay, several rods further west, and 

 at a considerably lower level. (See Plate IV.) 



5. On Agawam River. 



I have traced out three examples on this river of antediluvian date. One is in 

 Russell, on the west side of the present stream. The old bed is filled to a con- 

 siderable height with sand and gravel, compelling the river to find its way 

 through a rocky barrier. 



A second of these beds may be seen to the east of Chester village, at the junc- 

 tion of its east and west or principal branches. A third is some three miles above 

 this point on the east branch. (See my paper on Erosions.) 



I might refer to many other examples of ancient beds of rivers, not connected 

 with the Connecticut. But since most of these are older than the alluvial period, 

 they will more properly be noticed in my paper on Erosions. They would not be 

 mentioned here at all, were it not that the accumulation of detrital matter during 

 the last sojourn of the continent below the waters, seems to have been the means 

 of commencing many of those defiles in which rivers now run. 



Results, or Conclusions from the Facts. 



I shall now proceed to state the conclusions at which my own mind has arrived 

 from the facts which I have observed respecting surface geology, especially ter- 

 races, beaches, and drift. And as these conclusions are not based altogether 

 upon the details above given, I shall present a summary of the arguments by 

 which they are sustained, and the collateral facts and considerations on which 

 they rest. 



1. Postdiluvian terraces and beaches all lie above the coarse unstratified and 

 unmodified drift, as well as above the strise, furrows, and roclies moutonnes, con- 

 nected with drift. Hence the terraces and beaches are the result of operations 

 subsequent to the drift period. 



I wish not to be understood as maintaining that no genuine drift shows evidence 

 of stratification and other modifying effects of water. Such effects do present 

 themselves sometimes in the midst of detritus, which generally, in position and 

 character, affords unequivocal evidence of being true drift. Limited beds of sand 

 and clay are met with sometimes in the midst of such materials, and sometimes 

 we find masses of coarse irregular detritus and scattered blocks above deposits 

 that are distinctly sorted and stratified. But, as a general fact, the sorted and 

 stratified materials lie above the drift. 



I wish, also, to add, that it is no easy matter always to draw a line between 

 1 



