1892.] 
493 
[Davis. 
tion then arises, how could an esker thus left perched up in a 
channel on the ice be transferred to the surface of the ground 
below without interrupting its continuity? How could the con- 
tinuous sharp-ridged, single-crested esker at Auburndale, or its 
fellow at Newtonville, have been transferred from the ice to 
the ground? 
If the ice had melted away, we should expect that part beneath 
the esker gravels to have remained longest ; and from having 
lain in a channel when forming it must have come to rest on a 
ridge of ice. From such a ridge it would fall, part on one side, 
part on the other ; and its continuity would be completely lost. 
The only alternative to this suggestion is one made by Upham, to 
the effect that the percolating stream in the ice channel beneath the 
esker gravels would melt the ice on which it crept, and thus 
gradually lower the gravels to the ground without interrupting 
their continuity. Admitting that this is a possible process, it 
seems to me quite inconsistent with the open-work structure so 
common in the esker gravel. If the esker had been subjected to 
a gradual settlement, and had been traversed by currents of suf- 
ficient volume to melt the ice beneath, the open-work structure 
must have been filled up with sand and clay. More than this, 
such a process of gradual settlement must have greatly disturbed 
the bedded parts of the esker, much more than they are now 
found to be disturbed. For while slips and settlings are com- 
mon enough, they do not indicate a gradual settlement of one 
part after another, but settlement in larger masses, as will be re- 
ferred to later. 
Questions of this kind must often be settled as much by an 
agreement of various probabilities as by more strict method of 
demonstration ; and in this case, it seems to me to strain the 
probability severely to exjfiain the deposition of the Auburndale 
esker by the melting of the ice beneath it, or by any process that 
involves its gradual settlement from an ice foundation down to 
the ground. 
15 . ORIGIN OF ESKERS IN SUBGLACIAL CHANNELS. 
The alternative is that the esker was formed by a tumultuous 
subglacial stream where it entered a body of standing water ; a 
delta growing at the point of escape, and the stream rising from 
