82 TJie American Geologist. February, i89a 
recognized and criteria presented for distinguishing terraces 
of this origin from the slopes of beaches, fault-scarps, and 
river cut terraces. Davis* next applied the method to the de- 
limitation of the small esker-fan in Newtonville, Mass. In al! 
this work, the essential thing is the recognition of a topo- 
graphic feature which marks the former contact of the gravels 
and sands with now vanished masses of ice. 
"In a physical sense, we may speak of that feature in the 
topography of the glacial gravels which delimits or outlines 
vanished masses of ice as the ice-contact. The ice-contact is, 
in its most typical form, an even slope, with materials at or 
near the angle of repose. In this shape, it is frequently seen 
flanking one or both sides of an esker. the heads of sand-plains 
and the sides of many glacial lakelets or kettle-holes.f The 
ice-contact may arise in the unstratified group, as on the inner 
edge of boulder-belts ;;|; and the inner slope of frontal moraines 
is, almost without exception, where the ice has retreated from 
them, an ice-contact. 
The evenness of slope of the ice-contact may be perturbed 
by irregularities in the wall of the ice against ^vhich deposi- 
tion takes places, as by the presence of water-worn marginal 
crevasses and tunnels ; or by the slope of the ice-front so that 
deposition overlaps the edge of the ice. When such an ice- 
front melts out, the ice-contact will reflect the conditions of 
deposition of the detritus, in hummocky or kame-Hke slopes 
whose general declivity may be much less than that of the talus. 
Projecting, esker-like masses of gravel may be traced into the 
ice-field, and in the case of thick w^ater-laid deposits the head 
of the sand-plain may break down into a typical kame belt. 
Where the edge of the ice-sheet is deeply channeled, as 
seems to havC been the case in the last glacial epoch in New 
England, there is often a wide field in which gravels and sand:< 
were sheeted over the attenuated edge of the ice or about ice 
masses of unequal elevation, so that the ice contact is n(7 
longer a line on the map but must be represented as splitting 
with an inner edge on the iceward side of the kame belt or 
*Bull. Geo!. Soc. Am., Vol. i, 1890, pp. 195-202. 
t Warren Upham: Walden. Cochituate. and other lakes inclosed 
by modified drift, Proc. Boston Soc. Nat. Hist., xxv, 1891, 228-242. 
J J. B. Woodworth and C. F. Marbut: The Queen's River Mor- 
aine in Rhode Island, Journal of Geology (Chicago), iv, 1896, 698. 
