250 J. A. UDDEN 
tion indicated. Still this suggests itself as an alternative 
hypothesis, and it should be kept in mind, pending further 
investigation. The trends of the flutings are by no means more 
variable than the trends of ice scorings produced by glacial 
movements (see Fig. 2). 
To account for such a general motion, which conceivably may 
have been very small, two different hypotheses suggest them- 
selves: (1) Tundra conditions may have prevailed. On a slop- 
rd hy 
Fic. 2.— Map of Pottawattamie county, lowa, showing the locations and the 
directions of flutings observed in the loess. 
ing plain the annual temperature changes in a frozen tundra land 
would be apt to produce extensive creeping in the direction of 
least resistance, as towarda river. This would, no doubt, result 
in differential motion near the base of the frozen ground. (2) 
There may have been glacial conditions. With a sudden onset 
of arctic climate the area of accumulation of snow might extend 
far out beyond the margin of theice pushing out from the region 
of greatest accumulation, and were the most severe climate not 
of too long duration, the main continental glacier might come 
short of extending over all of such extra-marginal and perennial 
fields of snow. There might then be an extra-morainic névé. 
Perhaps a certain distribution of precipitation might favor such 
