>893.] 
489 
[D avis 
as they shifted from place to place, a convex lobate outline 
would be developed. The size of the delta would depend on 
the activity of the feeding stream from the ice, and on the dura- 
tion of the special conditions of its action. So complete a cor- 
respondence leaves no room for doubt. W e may therefore turn 
from this well-established foundation to the next step in the 
investigation : namely, the explanation of the gravel ridges or 
eskers so often extending backward from the head of the delta 
plateau. 
10 . THE FEEDING ESKERS. 
The occasional sections of the eskers .show that they consist of 
essentially the same materials as those found in the plateaus. The 
stones are similarly somewhat rounded but not well water-worn ; 
they are sometimes poorly stratified, but more commonly lie in 
much disorder ; although whether it should be asserted that they 
are absolutely unstratified is to me an open question. The distinc- 
tion between bad or imperfect stratification and an entire lack of 
stratification is difficult to draw. A frequent and characteristic 
feature of their structure is the occurrence of “openwork gravels,” 
as I have been accustomed to call them. The spaces between 
the pebbles are often left empty, although the layers adjoining 
contain plenty of fine material. The same structure has been 
seen in the gravel beds near the heads of certain sand plateaus. 
This I interpret as indicating hasty action, of whatever kind ; and 
I am disposed to regard it as an important link in the evidence 
leading to the explanation of the origin of eskers, as will appear 
hereafter. In several cases, it is possible to follow along an esker 
from its first appearance until it joins a sand plateau ; and I have 
met only one observer, a foreigner, experienced in glacial studies, 
who could avoid the conclusion that the esker and the sand pla- 
teau thus associated were formed contemporaneously. Perhaps 
it is pertinent to add that this observer was committed to the idea 
that eskers are formed in channels on the surface of the ice-sheet, 
and that he was frank enough to admit that, if the contempora- 
neous origin of the esker and the sand plateau were granted, the 
surface origin of the esker could not be maintained. I mention 
this the more freely because this well-known investigator has 
given what appear to me good reasons for thinking that some 
eskers are formed in surface channels ; but it does not appear 
