700 
GEOGRAPHY: W. M. DAVIS 
somewhat less than that of the cliff tops, gives the best indication of the 
local height of the ice margin. 
The bare hill sides of the upper valleys of the Clark fork drainage 
system show many faintly marked shoreKnes, up to altitudes of 4200 
feet; 20 or 24 such lines may be counted, one over the other, in some 
localities. These have been understood for some years past as recording 
the occurrence of a temporary lake of fluctuating level, to which the 
name of Lake Missoula has been given. Pardee pointed out in 1910 
that the lake must have resulted from the obstruction of Clark fork 
by the Canadian glacier at the head of Lake Pend Oreille,^ where marks 
of glacial scouring are recognizable in the steepening of the neighboring 
mountain spur on the south, between Lake Pend Oreille and upper 
FIG. 3. A STRONGLY TRUNCATED VALLEY-SIDE SPUR, ABOVE PARADISE, ^ MONTANA, 
LOOKING EAST: CLARK FORK IN FOREGROUND. 
Clark-fork valley, up to about the same altitude as that of the highest 
lake shoreline. The fluctuating level of the lake appears — following the 
explanation adopted in Sweden for similarly fluctuating glacial lakes — 
to result from the location of the outlet on the fluctuating surface of the 
glacier where it impinged on the mountain spur that divided its southern 
and southeastern branches. It is to be expected that various signs of 
rushing water should be found on the steepened slope of this spur; 
unfortunately I had no opportunity of examining that point during my 
visit of 1913. 
Now as the lake shorelines seem to prove that the upper tributary 
valleys of the Clark fork system were occupied by a lake while the main 
valley was invaded by the southeastern branch of the Kootenay-Pend 
