322 The Amt'fican Geologist. May.isbs 
the melting of an immense sheet of ice. The deposits thus had the same 
origin with the glacial drift ; but they have been f/iO(ii/iecf,he'mg separ- 
ated from the coarser portions, and further pulverized or rounded, and 
assorted in layers, by water.* 
According to such definition, and constantly recognizing 
the progressive deposition of both glacial and modified drift, 
with extension of their areas, as fast as the ice boundary re- 
treated, these terms are employed throughout my reports for 
the Geological Survey of Minnesota (1879 to 1888. and again 
in 1893 and 1894), on the glacial lake Agassiz for the Geolog- 
ical Survey of Canada (1890) and for the United States Geolog- 
ical Survey (Monograph XXV^ 1895), and in numerous pa- 
pers published by the Geological Society of America. 
Without attempting here a detailed description of the var- 
ious deposits or types of modified drift, dependent on diverse 
conditions of deposition, we may briefly characterize them as 
including (i) valley drift, related to present courses of drain- 
age, with which I class the chief development of the loess in 
the Missouri and Mississippi basins; (2) gravel, sand, and finer 
silt plains, not related to present streams, but representing 
areas of fluvial and lacustrine deposition when the part of the 
ice-sheet adjoining them was being melted away; (3) sand and 
gravel plateaus, and the loess paha of Iowa, more distinctly re- 
cording the ice borders against which they were amassed; (4) 
eskers, growing in length as the ice receded, being deposited 
in channels of ice-walled streams at and near their mouths; (5) 
kames, differing from eskers in the short duration of their par- 
ent glacial streams, best developed in marginal moraine belts; 
and (6) water-laid beds intercalated with till in drift sheets, 
drumlins, and marginal moraines. 
The volume of the modified drift in the Xew England 
states and westward to Minnesota, the Dakotas, and Manitoba, 
according to mv observations and estimates, may average vary- 
ingly a quarter to a half as much as the volume of the till. 
It was almost wholly supplied, with a considerable part of the 
till, as I believe, from drift that continued to be englacial or in- 
traglacial until ablation of the ice gave it over to rain and 
stream erosion, transportation, and deposition. 
♦Geology of New Hampshire, vol. iii. 1878, pp. 3, 4. 
