Origin of Eskcrs. — Crosby. 15 
Crevasses are the product of tensile stresses; and it is 
obvious that as, during the progressive cessation of its flow 
from' the south northward, each portion or zone of the ice be- 
comes stationary, the ice still in motion immediately to the 
northward will crowd focibly against it and tend to obliterate 
by compressive stresses any crevasses wdiich may interrupt its 
continuity. Overriding, or even a local thickening of the ice, 
may not always accompany the cessation of flow ; but it is cer- 
tainly difficult to see how, in general, crevasses could survive 
the progressive loss of motion. We have a magnificent illus- 
tration of this process in the Malaspina glacier, which is prac- 
tically free from crevasses and well endowed with persistent 
superglacial streams on its outer, drift-covered, marginal zone, 
although crevasses are a common feature of all the central and 
northern part of this great piedmont glacier, or at least com- 
mon enough to prevent the development of any important 
superglacial streams ; and we can also suppose that they are 
closed by pressure from the northward or clogged by super- 
ficial drift as the ice gradually ceases to flow. In fact, Rus- 
sell says that many of the crevasses are filled with clear, blue 
water ; and that they appear to be the scars left by rents in the 
tributary ice streams. Concerning the drainage of the Malas- 
pina glacier. Stone says :* "For some reason the glacial 
streams have either formed no subglacial tunnels under a mar- 
ginal zone of uncertain breadth, or the original tunnels have 
become blocked by ice or sediment or moraines so that the 
streams have been forced to form englacial tunnels, which be- 
come isuperglacial by the melting away of the overlying ice, 
and the streams continue such as they flow down the terminal 
ice slope. If the glacier continues to retreat, it seems probable 
that a ridge or series of ridges such as are now forming, and 
abandoned channels of these rivers, will be prolonged north- 
ward as far as the englacial channels reach. This furnishes 
an observational basis for the conclusion that during the re- 
treat of the ice-sheet, wherever the ice was very stagnant and 
the subglacial streams found their tunnels choked near their 
outlets, they freely rose into englacial or superglacial chan- 
nels." This is important testimony and all the more interest- 
ing as coming from a subglacialist. Surely, if a piedmont gla- 
cier on a narrow, sloping coastal plain, at the base of lofty 
• U. S. G. S., Mon., 34, pp. 421-422. 
