PHYSIOGRAPHIC CONDITIONS OF SAND-PLAIN FORMATION. 201 



tion, indicates that here also much more detritus was carried along than was 

 laid down. The sand-plain front was the goal at which most of the detritus 

 stopped, and hence its rapid growth. The clay beds that we should expect 

 to find as the final deposits of the glacial streams probably occupy the 

 meadow bottoms in front of the sand plains ; but as yet no sections clearly 

 manifesting the relation of the clay to the sand plains have been found. 



Sand Plains generally formed in local Bodies of Fresh Water. — Near the 

 coast, and up to an elevation of fifty or a hundred feet above present sea 

 level, in eastern Massachusetts, the water in which the sand plains were built 

 appears to have been ocean water; but the amount of submergence thus sur- 

 mised has not yet been fully worked out. Further inland, where plains are 

 found up to altitudes of a thousand or more feet above sea level, I think 

 the water in which they accumulated w^as fresh water, temporarily ponded 

 by the ice front. The reasons for this opinion are as follows: 



The ice of the last glacial epoch appears to have melted off of the country 

 first in the southern and later in the northern part of its area, producing a 

 general northward migration of the locus of sand-plain formation. Accepting 

 the generally current idea that the depression of the land diminished as the ice 

 retreated, it follows that the sand plains of later date should be of less eleva- 

 tion above present sea level than the earlier ones, if they were all deposited 

 in ocean water ; and this is not the fact. The sand plains of the interior 

 and northern part of New England, which must have been built at a rela- 

 tively late stage of ice melting, are of distinctly greater elevation above 

 existing sea level than those near the coast, which must have been built at 

 an earlier date. The interior sand plains are therefore regarded as having 

 been accumulated in local and temporary ponds, determined by the ever- 

 changing relation of the rock and drift topography to the frontal margin of 

 the retreating ice. Otherwise it would be necessary to suppose that the 

 submergence of the land increased as the ice melted away ; and while this 

 is manifestly not to be regarded as geologically impossible, it does not appear 

 to be accordant with the general results of glacial study thus far obtained. 



Relation of Sand Plains to other Glacial Deposits. — The relation of glacial 

 sand plains to two other similar forms of late or post-glacial deposits may be 

 briefly mentioned. In many cases the streams from the ice ran down open 

 valleys, and not into ponds of standing water. In such cases the valleys 

 were commonly clogged with flood-plain deposits of sand and gravel, often 

 of great extent. These are unlike the glacial sand plains in having no defi- 

 nite frontal slope, and hence in wanting also the steep-dipping fore-set beds, 

 of which the frontal slope is the external expression. The flood-plains are 

 indeed merely extended illustrations of what I have called the top-set beds 

 of the sand plains; but their connection with the back-set beds, which theory 

 leads me to suppose must exist, has not been traced out. The original flood- 

 plain, now the upper terrace, of the Merrimac, as described by Upham, in 



