Drift Mounds near Olympia^ Wash. — Rogers. 397 
gravel into which the wells extend from two to four feet, where 
permanent water is secured. 
Imagine this large tumuli plain to be covered with a sheet of ice 
reduced to a moderate thickness by ablation on the border of the 
retiring active glacier, and having, like the forest-covered border of 
the Malaspina ice-sheet in Alaska, no motion except the gradual 
waste b}^ liquefaction. However, the forces which modify the 
drift are yet active, so that boulders, cobblestones, gravel, and 
sand are being brought forward from the higher levels by periodic 
floods, the product of accelerated melting of the glacier during a 
prolonged increased temperature. From time to time, during the 
especially rapid melting in the summers, the greater floods would 
overflow the entire ice-plain, carrying along with them the smaller 
boulders, cobbles, gravel, sand, and clay or finest silt of the drift, 
which would be spread on the ice surface. It is a well observed 
fact that coarse gravel carried onto a plain by a strong flow of 
water has a tendency to accumulate in masses or heaps. A slight 
depression in the surface, or some obstruction, may arrest the on- 
ward movement of a boulder or large cobblestone, while the finer 
material would more or less pass on. This boulder might stop 
another, and so on, the general result being a somewhat unequal 
distribution of the drift over the surface of the ice. The flood as 
its force decreased would manifestly leave the whole surface more 
or less covered with the drift debris. As the sun sends its rays 
onto the drift thus distributed with its substratum of ice, three 
agencies modifying the drift claim our consideration, namel}', heat, 
gravitation, and the accumulation of water in the depressions. 
The thicker masses of debris would collect more of the heat of the 
sun and this would cause a faster melting of the ice or in other 
words would make a depression at such points, into which gravita- 
tion would cany additional drift by sliding from the adjacent ice 
surface and by the washing action of rains and streamlets. 
After formulating this hypothesis in regard to the tumuli of the 
" mound prairies," I was ver}^ much gratifled in finding the fol- 
lowing statements which add greatly to the strength of m}- argu- 
ment. 
Prof. I. C .Russell, in his work on the glaciers of Alaska,* 
gives on page 120 the following description of lakelets on the 
*"An Expedition to Mount St. Elias, Alaska," National Geographic 
Magazine, vol. iii, pp. .53-203, with nineteen plates, May 29, 1891. 
