ANCIENT WATER LEVELS OP CHAMPLAIN-HUDSON VALLEYS 185 



tion uninterrupted by dust and sand falls not only because of the 

 anchoring of the sands by ice over the dry land but also by reason 

 of the ice covering of the lakes or estuaries in which the clays 

 were dei)Osited. 



In this view, the sand partings of clays in this region should be 



thickest on the western side of the clay area and should wedge 



out to thin layers on the east, due allowance being made for the 



drifting, by currents, of the dust which falls into the water. 



The interpolation of sand partings by recurrent wind action 



p in something like cycles of one week agrees more closely with 

 the probable rate of deposition of the observed strata than the 

 supposition that the alternations depend on seasonal or diurnal 

 changes; and instead of allowing 5000 years for the deposition 

 of the clays in the Connecticut valley, for instance not more than 

 one 50th part of that time would probably suffice under the con- 

 ditions of excessive discharge of rock-flour from the neighboring 

 melting ice sheet. 



The sandy partings in clay often simulate the loess in character 

 and it is in them also that the equivalents of the " loess pupchen " 

 or " clay dogs " are frequently found. There is good evidence that 

 miany areas of loess are of eolian origin; but the sandy partings 

 in subaqueous clay areas differ from loess in that the sand has 

 come to rest beneath a water body rather than on an open air 

 surface. 



ik^ Succession of glacial clays. It has already been pointed out in 

 the chapter on the effects of retreating glaciers how deposits 

 on the same stratigraphic plane may be of different ages or stages 

 of glacial retreat. Each proglacial delta has its supplement in 

 clays extending from beneath it over the low ground in front of it. 

 Thus the fact that in the lower Hudson valley we find clays under- 

 lying gravels and sands does not show that there was first a 

 time of clay deposition followed by one of coarser deposits unless 

 it can also be shown that the gravels and sands were simul- 

 taneously deposited by normal streams. The ice contacts in that 

 region point clearly to a succession from north to south as the 

 ice front receded. 

 Exceptional reasons for predominance of clay deposits in the 



\ glacial series of the Hudson valley. The dominance of clays in 

 the Hudson valley from the Highlands to the mountains which 



