'894-] 219 [Woodworth. 
Shaler requires no further action for the production of eskers. 
It seems to tlie writer that most of the eskers which occur in 
soutliern New England are best explained on the supposition of 
subglacial channels filled with coarse detritus during the closing 
stages of the ice-sheet, even at so hite a time as when the tunnels 
were exposed to the incursion of detritus from point to point 
wliere the roof had been melted through. In this manner, it 
seems possible to explain the diversity in gradation of materials 
in the esker, and the occasional high j)oints or hummocks which 
cannot be explained by the methods proposed in the first part of 
this jjaper. Eskers were forming in the ice when terraces and 
sand-plains were forming outside of it, and often when the two 
series of deposits are so associated, there is a certain uniformity 
of level strongly suggestive of the melting or falling in of the 
arch pari passu witli the construction of the deposit in the 
channel. In other instances this appears not to have ha{»pened. 
Conclusions. 
1. The diversity of materials, structure, and shape of eskers 
shows that the term esker (osars, serpent kames) is applied in 
coiutnon usage to deposits having at least slightly different modes 
of origin. 2. It follows from this, that each esker should be 
diagnosed upon its own merits, with regard to its external and 
internal structure, and its origin. 3. The diversity in the slopes, 
crest-line, and course of the same esker demands careful interpre- 
tation as to the mode of deposition and relations to the ice-sheet 
held by particular segments of the esker. 4. The steep slopes of 
some eskers indicate an adjustment to gravity upon the liquefac- 
tion of supporting ice-walls. 5. It follows from the limitations 
in cross-section of eskers, that where the deposit is now high and 
low, it must in places have been originally wide and narrow 
respectively. 6. In some eskers, maxima of change in direction 
correspond in moment with maxima of change in elevation of the 
crest-line. 7. So far as eskers are subglacial, they reflect condi- 
tions resident in the bottom of the ice-sheet. 8. The limitations 
in the cross-section of eskers demand some limiting agent in the 
ice-sheet, and the ice-arch alone fulfils the requirements. 9. 
Eskers tend to lie in meridional valleys, and to lie on one side of 
