386 PROCEEDINGS: BOSTON SOCIETY NATURAL HISTORY 
Other words, the ice may flow by intergranular melting and freezing 
without any excess of melting over freezing. 
Again, it is probable that during the period of glacial erosion the 
basal pressure is sufficient to prevent the development of both 
tunnels and crevasses. Nansen found neither crevasses nor super¬ 
glacial streams nor evidence of appreciable surface melting; and it 
is certain that surface crevasses would be closed at no great depth 
by the freezing of water in them, as long as the mean annual tem¬ 
perature remains, as now, far below the freezing point. We must 
conclude, therefore, that glacial erosion was, and is now except in 
the case of alpine glaciers whose valleys have been swept free of 
debris, accomplished in the practical absence of w^ater, derived 
either from the basal or the superficial melting of the ice. 
We thus find that, even in the case of a waning ice cap, supergla¬ 
cial and still more narrowlv subsclacial drainage is confined to the 
marginal portion of the ice, and in view of the observations by 
Nansen and others on the inland ice of a mountainous country, we 
may question whether crevasses could have been a prominent feature 
of even the marginal portion of the Pleistocene ice sheet on the 
plain portions of the glaciated area. Or, in other words, is it not 
probable that on an approximately level but rough surface, remote 
from mountainous tracts, the ice Avould, in general, have ceased to 
flow before it became thin enough to permit effective crevassing or 
the development of fissures traversing its entire thickness ? I am 
inclined to think that the burden of proof rests upon those who 
assume the negative, although glacial potholes are commonly sup¬ 
posed to testify to the existence of at least occasional crevasses, and 
inequality of basal melting may be cited as a possible cause of cre¬ 
vassing not dependent upon flow. 
Stone says (’99, p. 240), “Neglecting basal melting, we divide the 
ice sheet into a zone or area of diffused superficial waters, a zone of 
superficial streams, and a zone of subglacial streams.” I accept these 
zones, but hold farther that, in large jjart, the second zone extended 
to the margin of the ice, the third or outer zone being characterized 
by superglacial streams above and subglacial streams below, com¬ 
peting for the englacial drift, of Avhich eskers must be chiefly formed. 
Undoubtedly, the very latest direct Avork of the ice sheet, before 
its motion finally ceased, Avas the building of the drumlins. In the 
lee sides of a few drumlins a limited amount of gravel is interstrat- 
