Eskers of the Kunsan Epoch. — Uershey. 241 
that the ice was in motion during the formation of the strati- 
fied drift, but not enough to prove that these deposits were 
generally overridden by it. Now, we cannot suppose that the 
tunnels remained intact and stationarj'- throughout a distance 
of twenty or more miles while the ice continued to move for- 
ward. As the ice undoubtedly did move forward (most 
decisive evidence of this movement will be produced 
when I come to discuss the transported rock masses), and 
as also the deposits at the east end of the belt show no 
more evidence of disturbance than those at the west end. 
which is practically none, we may conclude that the ice-front 
progressively retreated eastward across the district, and that 
the deposits as now seen are only those which were laid down 
about the mouths of the tunnels. Thus the analogy between 
the systems of " belts" and a subaerial river system is more 
apparent than real. However, as the situation of the deposits 
was determined by the position of the tunnels, which position 
was partly dependent on the topography of the district gla- 
ciated, the belts as they now appear were outlined by the tun- 
nels, and consequently we may judge of the courses of the 
subglacial streams by a study of the belts. 
When the retreating ice-front halted for a time along some 
line, such halt being similar to those which, in a later glacial 
epoch, formed the terminal moraines, the deposits made at the 
mouths of the tunnels during a number of summers closely 
adjoined or overlaid each other, so as to form more prominent 
accumulations than when the mouths of the tunnels were in 
rapid recession. This is my explanation of the so-called 
" special areas " of these gravel deposits. Every one of them 
indicates a halt of some length in the recession of the ice- 
margin. In other words, they correspond to the kames, or 
short ridges of diagonally stratified gravel, which are often 
found in depressions or gaps in the terminal moraines. 
Let us follow a given ridge, as, for example, one of the many 
which together constitute the Pecatonica belt, and study it in 
more detail than we have yet done. Traversing it from its 
eastern end, we see it gradually grow larger until, at one, two, 
or three miles from its point of beginning, it is suddenly de- 
veloped into a "special area,"' and then just as suddenly it 
may die out completely. One of these short ridges constitutes 
