12 
Journal of Agricultural Research 
Vol. XI, No. i 
dry sites are very unfavorable to the establishment of seedlings even though 
germination may take place. Furthermore, irregularity in the areas of 
young growth occurred on all sorts of sites. This could lead to but one 
conclusion: wherever ground fire occurred no reproduction appeared, 
except close to seed trees where seed could be cast upon the burned ground 
after the fire. 
From this it was but a step to the complete explanation: wherever the 
duff and litter were burned out of the forest floor, there developed an 
area barren of reproduction; wherever the duff and litter were not burned 
out of the forest floor, there developed an area of more or less dense 
reproduction. Therefore, the duff must be the controlling element: the 
duff must be the storage medium of the seed, and that seed must have 
been produced and stored in the forest floor before the fire and have 
retained its viability through the fire. 
Before this conclusion is accepted, however, another possible source 
of seed must be considered. Is it not possible that cones carried through 
the fire on the crowns of trees severely burned or killed furnished the seed 
from which the young growth originated ? After the fire these cones may 
possibly have opened and dispersed their seed, becoming in that way 
an overhead source for the restocking of the burn. In fact, a very small 
percentage of germination of white-pine, noble-fir, and cedar seed has 
been secured from seed which passed through a crown fire. But even 
though this source does contribute some seed, it does not explain the great 
mass of reproduction, which, by its mosaic occurrence, demonstrates 
conclusively the impossibility of its having come from overhead seed 
distribution subsequent to the fire. The principal factor in reproduc¬ 
tion after* fire must be the seed stored in the duff. 
UPPER CISPUS BURN, A SINGLE BURN ON WHICH THE DUFF WAS COM¬ 
PLETELY DESTROYED BY THE FIRE 
In an area on the Upper Cispus River, Rainier National Forest, burned 
in 1910, there was no reproduction except in the vicinity of seed trees; 
and the seedlings were more numerous close to the edge of blocks of seed 
trees than at a distance, which is invariably the condition where migra¬ 
tion occurs from wind-blown seed. This is a typical example of an area 
made barren by a single fire which completely destroys the duff and 
thereby precludes the possibility of reproduction from seed stored in the 
forest floor. The complete destruction of the duff may have been due 
not so much to an unusually intense fire as to the extreme lightness of the 
dry volcanic soil of the region, which is easily “burned out” by ground 
fire. The conditions found on this area corrorborate the conclusions 
reached in the study of the Columbia burn. Where the duff is com¬ 
pletely burned, reproduction is found only in inverse numerical proportion 
to the distance from seed trees. 
