189a.] 
491 
[Davis. 
must have been required for its deposition. But if that had been 
the case, the margin of the stagnant ice must have been melted 
by a significant amount while the delta was growing, and thus 
have opened a space at the head of the delta to be filled up with 
gravel as fast as the ice melted back. No deposits that can be 
interpreted as formed in such a manner have been found, if we 
except certain small slanting layers, insignificant when compared 
to the mass of the plateau. Although two well-exposed sections at 
the head of a sand plateau have been attentively examined, 
they do not indicate a recession of the ice of more than ten or 
twenty feet, while the forward growth of the plateau extended 
over several hundred feet. It may be that the ice melting was 
very slow ; but it appears more probable that the delta growth 
was relatively rapid, the product of gushing streams heavily laden 
with sand and gravel. This view is borne out by the structure 
of the open-work gravels above mentioned in the eskers and in 
the coarser beds near the head of the plateau. 
12 . LOCAL AND SPASMODIC GROWTH OF ESKERS AND SAND 
PLATEAUS. 
A new question now arises. If the forward delta growth was * 
so much more rapid than the backward melting of the ice, why 
was not the whole country covered with sand plateaus? At 
Auburndale the sand plateau grew forward at least twenty times 
faster than the ice melted backward. The ice has melted back- 
ward all across New England. Hence unless some other factor 
is to be considered, the sand plateau should extend about six 
thousand miles southward. This is of course simply a reductio ad 
absurdum , in order to emphasize the conclusion that the rapid 
growth of the sand delta was a brief, spasmodic, temporary affair, e 
merely a local short-lived episode in the deliberate recession 
of the ice. The sand plateau therefore marks the transient out- 
let of a glacial stream, well charged with gravel and sand, and 
here flowing from the ice into a body of standing water. It might 
perhaps be suggested that the stream had a permanent outlet 
along the Auburndale esker, and that it was only at times charged 
with detritus in sufficient quantities to build a sand plateau ; but 
this assumption would be contrary to the uniform experience of 
Arctic observers. Holst of Sweden, when in Greenland, and 
